CN112488246A - Information processing method and system based on network live broadcast and online e-commerce delivery - Google Patents

Abstract

According to the information processing method and system based on the network live broadcast and the online E-commerce delivery, the information identification degrees of a plurality of label information and the label similarity rates among the label information are determined according to the first information list and the second information list. And verifying the label information based on the information identification degree of the information and the label similarity rate between different label information, and counting the target label information passing the verification. And acquiring the message of the intelligent terminal and determining a time sequence continuity curve and a keyword queue of the message. And setting an access interception mechanism when judging that the intelligent terminal carries the Trojan horse program based on the time sequence continuity curve and the keyword sequence. Therefore, whether the intelligent terminal carries the Trojan program or not can be timely and accurately detected by carrying out time sequence continuity detection and natural language analysis on the interactive information of the intelligent terminal, so that the intelligent terminal carrying the Trojan program is accessed and intercepted, and the information safety of an e-commerce platform and other intelligent terminals is ensured.

Description

Information processing method and system based on network live broadcast and online e-commerce delivery

Technical Field

The disclosure relates to the technical field of network live broadcast and online e-commerce, in particular to an information processing method and system based on network live broadcast and online e-commerce delivery.

Background

With the development of the internet, the life style of people has changed greatly. Taking shopping as an example, online shopping is now the main way for people to shop. In order to increase the sales volume of products on line, the merchant usually communicates with the user on the e-commerce platform through a live network, so as to realize online goods delivery. The essence of communication between a merchant and a user is information interaction between an e-commerce platform and an intelligent terminal, and therefore, how to ensure the information security of the e-commerce platform and the intelligent terminal is a technical problem to be solved urgently at present.

Disclosure of Invention

In order to solve the technical problems in the related art, the present disclosure provides an information processing method and system based on live webcasting and online e-commerce delivery.

In a first aspect, an information processing method based on live webcasting and online e-commerce delivery is provided, and is applied to a cloud server communicating with a plurality of intelligent terminals, and the method at least comprises the following steps:

for each intelligent terminal, determining information identification degrees of a plurality of label information to be verified, which are used for determining the time sequence continuity of the interactive information and extracting keywords of the interactive information, and label similarity rates among different label information based on a first information list and a second information list which are extracted and used for recording the interactive information of the intelligent terminal;

verifying the plurality of label information based on the determined information identification degrees of the plurality of label information and the label similarity rates among different label information, and counting at least three target label information passing the verification; the information identification degree of the target label information is greater than the set identification degree, and the label similarity rate between the target label information is smaller than the set similarity rate;

acquiring a message of the intelligent terminal in a current time period determined based on an operation log of the cloud server, generating a time sequence continuity curve of the message in the current time period according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate among the target label information, and extracting a keyword queue of the message in the current time period;

judging whether the intelligent terminal carries a Trojan program or not based on the time sequence continuity curve and the keyword sequence; if so, setting an access interception mechanism for intercepting the intelligent terminal according to the terminal configuration information of the intelligent terminal.

Optionally, determining whether the intelligent terminal carries a trojan program based on the time sequence continuity curve and the keyword sequence includes:

obtaining a critical curve segment in the time sequence continuity curve, wherein the critical curve segment comprises a curve segment with a curve slope larger than a critical point of a set slope; equally dividing the critical curve segment into at least two sub-line segments; calculating the curve slope change coefficient included in each sub-line segment of the at least two sub-line segments in parallel; mapping each calculated curve slope change coefficient to a preset value interval and calculating the weighted sum of the mapping values of each curve slope change coefficient in the preset value interval;

when the weighted sum is larger than a target value, extracting a word vector of each keyword in the keyword sequence; clustering the extracted multiple word vectors based on curve features corresponding to the time sequence continuity curves to obtain at least two clustering sets; calculating the feature difference degree between the word vectors in each cluster set;

when the difference value of each feature difference degree and the set difference degree is smaller than a preset threshold value, selecting a cluster set corresponding to the maximum feature difference degree as a detection set, determining a word vector matrix corresponding to the detection set, and searching whether a target word vector matrix with the matrix similarity between the word vector matrix and the word vector matrix larger than the set similarity exists or not from a preset database; if yes, judging that the intelligent terminal carries the Trojan horse program; and matrix elements in the target word vector matrix are elements corresponding to abnormal keywords.

Optionally, the method further includes clustering the extracted multiple word vectors based on curve features corresponding to the time series continuity curve to obtain at least two cluster sets, and calculating a feature difference between the word vectors in each cluster set, and further includes:

determining the characteristic dimension quantity of the curve characteristic, normalizing the position information of each characteristic value in the characteristic value queue corresponding to the curve characteristic to obtain a relative position coefficient corresponding to each characteristic value, correcting each characteristic value by using the relative position coefficient to obtain a correction value corresponding to each characteristic value, and calculating a characteristic configuration parameter corresponding to the curve characteristic according to the correction value; the characteristic configuration parameters are used for updating model parameters of a preset clustering model, and the preset model parameters are kmeans mean clustering models;

extracting current model parameters of the preset clustering model, and calculating the current clustering accuracy of the preset clustering model according to a correlation coefficient between the current model parameters and the feature configuration parameters; when the current clustering accuracy is lower than a set accuracy, updating the preset clustering model by using the characteristic configuration parameters;

determining a clustering feature number based on the updated preset clustering model, and judging whether the vector dimension number of the word vector is equal to the clustering feature number; if not, adjusting the vector dimension number of the word vector based on the feature configuration parameters so as to enable the vector dimension number of the word vector to be equal to the clustering feature number;

inputting the adjusted word vectors into the updated preset clustering model and obtaining at least two clustering sets output by the updated preset clustering model; and aiming at each cluster set, calculating cosine distances between each word vector in the cluster set and other word vectors, screening the cosine distances according to the clustering weight of each word vector in the cluster set, and calculating the mean value of the cosine distances obtained by screening to obtain the feature difference.

Optionally, determining, based on the extracted first information list and second information list for recording the interactive information of the intelligent terminal, information identification degrees of a plurality of pieces of tag information to be verified, which are used for determining time sequence continuity of the interactive information and for extracting keywords of the interactive information, and tag similarity ratios between different pieces of tag information, further comprising:

listing a first information set used for representing communication records between the intelligent terminal and the cloud server in the first information list according to a time sequence, listing a second information set used for representing communication records between the intelligent terminal and other intelligent terminals in the second information list according to the time sequence, generating a first information flow track of the first information list based on the first information set, and generating a second information flow track of the second information list based on the second information set; the number of first track nodes in the first information flow track is different from the number of second track nodes in the second information flow track, and the first track nodes and the second track nodes have different node information dimensions;

acquiring a time sequence continuous track for representing the interactive information and target information of an information subject of the interactive information, which are represented by a node information dimension corresponding to one first track node in the first information flow track; determining a second track node with the largest node information dimension in the second information flow track as a target track node, and packaging the target information into the target track node according to a time sequence difference value between the first information flow track and the second information flow track so as to obtain configuration information corresponding to the target information in the target track node; determining a time sequence consistency weight between the first information list and the second information list according to the matching degree between the configuration information and the target information;

performing time sequence correction on the first information stream track and the second information stream track by adopting the time sequence consistency weight, and splicing the corrected first information stream track and the corrected second information stream track to obtain a third information stream track; extracting node description information corresponding to each third track node in the third information flow track and classifying the node description information corresponding to each third track node to obtain a first information group and a second information group;

determining a plurality of pieces of first label information to be verified for determining the time sequence continuity of the interactive information based on the node description information in the first information group, determining a plurality of pieces of second label information to be verified for determining the keyword for extracting the interactive information based on the node description information in the second information group, and screening out a plurality of pieces of third label information to be verified for determining the time sequence continuity of the interactive information and extracting the keyword for the interactive information according to the first label information and the second label information; mapping the label index value corresponding to each piece of third label information to a preset numerical value list to obtain the information identification degree corresponding to each piece of third label information; and calculating cosine distances between information vectors of every two pieces of third label information, and weighting the calculated cosine distances to obtain label similarity rates among the plurality of pieces of third label information.

Optionally, the classifying the node description information corresponding to each third trace node to obtain a first information packet and a second information packet specifically includes:

dividing a track curve corresponding to the third information stream track into a plurality of curve segments according to a preset time interval, calculating a curve slope for representing the track stability of each curve segment, and generating a track distribution list corresponding to the third information stream track according to the curve slope; classifying the node description information corresponding to each third track node for the first time based on the information association degree of the list information in the track distribution list to obtain a first initial group and a second initial group;

when it is determined that an information compatibility identifier exists between the node description information under the first initial grouping and the node description information under the second initial grouping, adjusting at least one node description information under the first initial grouping to be under the second initial grouping based on the determined information compatibility between the node description information under the first initial grouping and the node description information under the second initial grouping;

calculating a first information configuration coefficient of the node description information under the first initial grouping after the adjustment and a second information configuration coefficient of the node description information under the second initial grouping after the adjustment are completed, and determining a coefficient difference value between the first information configuration coefficient and the second information configuration coefficient; judging whether the coefficient difference is larger than a set difference or not; when the coefficient difference is larger than the set difference, determining a first initial packet after adjustment as the first information packet and determining a second initial packet after adjustment as the second information packet; and when the coefficient difference is smaller than or equal to the set difference, continuing to adjust at least one piece of node description information under the first initial grouping to the second initial grouping, and returning to the step of calculating a first information configuration coefficient of the adjusted node description information under the first initial grouping and a second information configuration coefficient of the adjusted node description information under the second initial grouping.

Optionally, the method further comprises:

acquiring a modification instruction for modifying the set identification degree and the set similarity rate;

and modifying the set identification degree and the set similarity according to the modification instruction.

In a second aspect, a cloud server is provided, the cloud server being in communication with a plurality of smart terminals, the cloud server being configured to at least:

for each intelligent terminal, determining information identification degrees of a plurality of label information to be verified, which are used for determining the time sequence continuity of the interactive information and extracting keywords of the interactive information, and label similarity rates among different label information based on a first information list and a second information list which are extracted and used for recording the interactive information of the intelligent terminal;

verifying the plurality of label information based on the determined information identification degrees of the plurality of label information and the label similarity rates among different label information, and counting at least three target label information passing the verification; the information identification degree of the target label information is greater than the set identification degree, and the label similarity rate between the target label information is smaller than the set similarity rate;

acquiring a message of the intelligent terminal in a current time period determined based on an operation log of the cloud server, generating a time sequence continuity curve of the message in the current time period according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate among the target label information, and extracting a keyword queue of the message in the current time period;

judging whether the intelligent terminal carries a Trojan program or not based on the time sequence continuity curve and the keyword sequence; if so, setting an access interception mechanism for intercepting the intelligent terminal according to the terminal configuration information of the intelligent terminal.

Optionally, the determining, by the cloud server, whether the intelligent terminal carries the trojan program based on the time sequence continuity curve and the keyword sequence specifically includes:

obtaining a critical curve segment in the time sequence continuity curve, wherein the critical curve segment comprises a curve segment with a curve slope larger than a critical point of a set slope; equally dividing the critical curve segment into at least two sub-line segments; calculating the curve slope change coefficient included in each sub-line segment of the at least two sub-line segments in parallel; mapping each calculated curve slope change coefficient to a preset value interval and calculating the weighted sum of the mapping values of each curve slope change coefficient in the preset value interval;

when the weighted sum is larger than a target value, extracting a word vector of each keyword in the keyword sequence; clustering the extracted multiple word vectors based on curve features corresponding to the time sequence continuity curves to obtain at least two clustering sets; calculating the feature difference degree between the word vectors in each cluster set;

when the difference value of each feature difference degree and the set difference degree is smaller than a preset threshold value, selecting a cluster set corresponding to the maximum feature difference degree as a detection set, determining a word vector matrix corresponding to the detection set, and searching whether a target word vector matrix with the matrix similarity between the word vector matrix and the word vector matrix larger than the set similarity exists or not from a preset database; if yes, judging that the intelligent terminal carries the Trojan horse program; and matrix elements in the target word vector matrix are elements corresponding to abnormal keywords.

In a third aspect, a cloud server is provided, which includes a processor and a memory, which are in communication with each other, and the processor retrieves a computer program from the memory and runs the computer program, so as to implement the above method.

In a fourth aspect, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when executed, implements the above-described method.

The technical scheme provided by the embodiment of the disclosure can have the following beneficial effects.

Firstly, the information identification degrees of a plurality of label information and the label similarity rate between different label information are determined according to the first information list and the second information list.

And secondly, verifying the plurality of label information based on the determined information identification degrees of the plurality of label information and the label similarity rate among different label information, and counting at least three target label information passing the verification.

And then, acquiring the message of the intelligent terminal in the current time period, generating a time sequence continuity curve of the message according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate among the target label information, and extracting a keyword queue of the message.

And finally, setting an access interception mechanism when the intelligent terminal is judged to carry the Trojan horse program based on the time sequence continuity curve and the keyword sequence.

Therefore, whether the intelligent terminal carries the Trojan program or not can be timely and accurately detected by carrying out time sequence continuity detection and natural language analysis on the interactive information of the intelligent terminal, so that the intelligent terminal carrying the Trojan program is accessed and intercepted, and the information safety of an e-commerce platform and other intelligent terminals is ensured.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.

Fig. 1 is a schematic flow chart of an information processing method based on live webcasting and online e-commerce delivery according to the present disclosure;

FIG. 2 is a block diagram illustrating an information processing apparatus based on live webcasting and online e-commerce delivery, according to an example embodiment;

FIG. 3 is an architectural diagram illustrating an information handling system based on live webcasting and online e-commerce delivery, according to an exemplary embodiment;

fig. 4 is a schematic diagram illustrating a hardware structure of a cloud server according to another exemplary embodiment.

Detailed Description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

The inventor analyzes the existing online e-commerce delivery technology, and finds that the existing e-commerce platform usually interfaces with a plurality of intelligent terminals, that is, the e-commerce platform and the plurality of intelligent terminals communicate information. Therefore, if one intelligent terminal carries the trojan program, the information security of the e-commerce platform and other intelligent terminals can be affected. Therefore, how to timely and accurately detect the trojan horse of the intelligent terminal is the key for ensuring the information security of the e-commerce platform and the intelligent terminal.

In order to solve the technical problems, embodiments of the present invention provide an information processing method and system based on live webcasting and online e-commerce delivery, which can detect whether an intelligent terminal carries a trojan program timely and accurately by performing time sequence continuity detection and natural language analysis on interactive information of the intelligent terminal, so as to access and intercept the intelligent terminal carrying the trojan program, and ensure information security of an e-commerce platform and other intelligent terminals.

To achieve the above object, please refer to fig. 1, which is a flowchart of an information processing method based on live webcasting and online e-commerce delivery according to an embodiment of the present invention, the method may be applied to a cloud server in communication with a plurality of intelligent terminals, where the cloud server is a server deployed in a cloud by an e-commerce platform, and further, the method may specifically include the contents described in the following steps S110 to S140.

Step S110, aiming at each intelligent terminal, based on the extracted first information list and the second information list for recording the interactive information of the intelligent terminal, determining the information identification degrees of a plurality of label information to be verified, which are used for determining the time sequence continuity of the interactive information and extracting the keywords of the interactive information, and the label similarity rates among different label information.

In step S110, the first information list is used to record interaction information (direct interaction information) between the intelligent terminal and the cloud server, and the second information list is used to record interaction information (indirect interaction information) between the intelligent terminal and other intelligent terminals.

Step S120, verifying the plurality of label information based on the determined information identification degrees of the plurality of label information and the label similarity rates among different label information, and counting at least three target label information passing the verification; the information identification degree of the target label information is larger than the set identification degree, and the label similarity rate between the target label information is smaller than the set similarity rate.

In step S120, the set recognition degree and the set similarity rate are set according to the number of the intelligent terminals, which is not limited herein.

Step S130, obtaining the message of the intelligent terminal in the current time period determined based on the operation log of the cloud server, generating a time sequence continuity curve of the message in the current time period according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate between the target label information, and extracting the keyword queue of the message in the current time period.

In step S130, the running log is used to record information processing records of the cloud server, the message messages include but are not limited to text messages, voice messages, and picture messages, the time sequence continuity curve is used to represent the reliability of the message messages in time sequence, and the keyword queue is used to record keywords in the message messages.

Step S140, judging whether the intelligent terminal carries a Trojan program or not based on the time sequence continuity curve and the keyword sequence; if so, setting an access interception mechanism for intercepting the intelligent terminal according to the terminal configuration information of the intelligent terminal.

In step S140, the access intercepting mechanism is configured to intercept information sent by the smart terminal to the cloud server and other smart terminals, and is further configured to intercept information sent by the cloud server and other smart terminals to the smart terminal.

By applying the contents described in the above steps S110 to S140, the information identification degrees of a plurality of tag information and the tag similarity rates between different tag information are first determined according to the first information list and the second information list, secondly, verifying the plurality of label information based on the determined information identification degrees of the plurality of label information and the label similarity rate between different label information and counting at least three target label information passing the verification, and then acquiring the message of the intelligent terminal in the current time period, generating a time sequence continuity curve of the message according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate between the target label information, extracting a keyword queue of the message, and finally setting an access interception mechanism when judging that the intelligent terminal carries the Trojan program based on the time sequence continuity curve and the keyword sequence. Therefore, whether the intelligent terminal carries the Trojan program or not can be timely and accurately detected by carrying out time sequence continuity detection and natural language analysis on the interactive information of the intelligent terminal, so that the intelligent terminal carrying the Trojan program is accessed and intercepted, and the information safety of an e-commerce platform and other intelligent terminals is ensured.

In practical application of the above scheme, the inventor finds that the determined tag information is defective when determining the tag information to be verified. The reason for this is because the consistency in time sequence between the first information list and the second information list is not considered. In order to improve the above technical problem and thus accurately and completely determine the tag information and the corresponding information identification degree and tag similarity ratio thereof, the method described in step S110 may further include determining the information identification degrees of a plurality of tag information to be verified, which are used for determining the time sequence continuity of the interactive information and for extracting the keyword of the interactive information, based on the first information list and the second information list extracted and used for recording the interactive information of the intelligent terminal, and the tag similarity ratio between different tag information, and further include the following steps S111 to S114.

Step S111, listing a first information set used for representing communication records between the intelligent terminal and the cloud server in the first information list according to a time sequence, listing a second information set used for representing communication records between the intelligent terminal and other intelligent terminals in the second information list according to a time sequence, generating a first information flow track of the first information list based on the first information set, and generating a second information flow track of the second information list based on the second information set; the number of first track nodes in the first information flow track is different from the number of second track nodes in the second information flow track, and the first track nodes and the second track nodes have different node information dimensions.

Step S112, acquiring a time sequence continuous track for representing the interactive information and target information of an information subject of the interactive information, which are represented by a node information dimension corresponding to one first track node in the first information flow track; determining a second track node with the largest node information dimension in the second information flow track as a target track node, and packaging the target information into the target track node according to a time sequence difference value between the first information flow track and the second information flow track so as to obtain configuration information corresponding to the target information in the target track node; and determining the time sequence consistency weight between the first information list and the second information list according to the matching degree between the configuration information and the target information.

Step S113, performing time sequence correction on the first information stream track and the second information stream track by adopting the time sequence consistency weight, and splicing the corrected first information stream track and the corrected second information stream track to obtain a third information stream track; and extracting node description information corresponding to each third track node in the third information flow track, and classifying the node description information corresponding to each third track node to obtain a first information group and a second information group.

Step S114, determining a plurality of pieces of first label information to be verified for determining the time sequence continuity of the interactive information based on the node description information in the first information group, determining a plurality of pieces of second label information to be verified for determining the keyword for extracting the interactive information based on the node description information in the second information group, and screening out a plurality of pieces of third label information to be verified for determining the time sequence continuity of the interactive information and extracting the keyword for the interactive information according to the first label information and the second label information; mapping the label index value corresponding to each piece of third label information to a preset numerical value list to obtain the information identification degree corresponding to each piece of third label information; and calculating cosine distances between information vectors of every two pieces of third label information, and weighting the calculated cosine distances to obtain label similarity rates among the plurality of pieces of third label information.

It can be understood that, through the steps S111 to S114, the consistency between the first information list and the second information list in time sequence can be analyzed, so as to avoid the absence of the determined tag information, and further accurately and completely determine the tag information, the information identification degree corresponding to the tag information, and the tag similarity.

On the basis, the classifying the node description information corresponding to each third trace node to obtain the first information packet and the second information packet described in step S113 may specifically include the contents described in the following steps S1131 to S1133.

Step S1131, dividing a track curve corresponding to the third information stream track into a plurality of curve segments according to a preset time interval, calculating a curve slope for representing the track stability of each curve segment, and generating a track distribution list corresponding to the third information stream track according to the curve slope; and classifying the node description information corresponding to each third track node for the first time based on the information association degree of the list information in the track distribution list to obtain a first initial group and a second initial group.

Step S1132, when it is determined that there is an information compatibility identifier between the node description information in the first initial group and the node description information in the second initial group, adjusting at least one node description information in the first initial group to the second initial group based on the determined information compatibility between the node description information in the first initial group and the node description information in the second initial group.

Step S1133, calculating a first information configuration coefficient of the node description information in the first initial group after the adjustment is completed and a second information configuration coefficient of the node description information in the second initial group after the adjustment is completed, and determining a coefficient difference between the first information configuration coefficient and the second information configuration coefficient; judging whether the coefficient difference is larger than a set difference or not; when the coefficient difference is larger than the set difference, determining a first initial packet after adjustment as the first information packet and determining a second initial packet after adjustment as the second information packet; and when the coefficient difference is smaller than or equal to the set difference, continuing to adjust at least one piece of node description information under the first initial grouping to the second initial grouping, and returning to the step of calculating a first information configuration coefficient of the adjusted node description information under the first initial grouping and a second information configuration coefficient of the adjusted node description information under the second initial grouping.

In specific implementation, by applying the steps S1131 to S1133, the node description information corresponding to the third trace node can be accurately classified, so that the first information packet and the second information packet are accurately obtained.

In a specific implementation process, in order to accurately and reliably determine whether the intelligent terminal carries the trojan program and avoid erroneous determination, the determination of whether the intelligent terminal carries the trojan program based on the time sequence continuity curve and the keyword sequence described in step S140 may further include the following contents described in steps S141 to S143.

Step S141, obtaining a critical curve segment in the time sequence continuity curve, wherein the critical curve segment is a curve segment comprising a critical point of which the curve slope is greater than a set slope; equally dividing the critical curve segment into at least two sub-line segments; calculating the curve slope change coefficient included in each sub-line segment of the at least two sub-line segments in parallel; and mapping each calculated curve slope change coefficient to a preset value interval and calculating the weighted sum of the mapping values of each curve slope change coefficient in the preset value interval.

Step S142, when the weighted sum is larger than the target value, extracting a word vector of each keyword in the keyword sequence; clustering the extracted multiple word vectors based on curve features corresponding to the time sequence continuity curve to obtain at least two clustering sets, and calculating feature difference between the word vectors in each clustering set.

Step S143, when the difference value between each feature difference degree and the set difference degree is smaller than a preset threshold value, selecting a cluster set corresponding to the maximum feature difference degree as a detection set, determining a word vector matrix corresponding to the detection set, and searching whether a target word vector matrix with the matrix similarity between the word vector matrix and the word vector matrix larger than the set similarity exists in a preset database; if yes, judging that the intelligent terminal carries the Trojan horse program; and matrix elements in the target word vector matrix are elements corresponding to abnormal keywords.

In practical application, based on the steps S141 to S143, whether the intelligent terminal carries the trojan program can be accurately and reliably determined.

In a more specific embodiment, the clustering, which is described in step S142, the extracted multiple word vectors based on the curve features corresponding to the time series continuity curve to obtain at least two cluster sets, and calculating the feature difference between the word vectors in each cluster set may specifically include the following contents described in step S1421 to step S1424.

Step S1421, determining the feature dimension number of the curve feature, normalizing the position information of each feature value in the feature value queue corresponding to the curve feature to obtain a relative position coefficient corresponding to each feature value, correcting each feature value by using the relative position coefficient to obtain a correction value corresponding to each feature value, and calculating a feature configuration parameter corresponding to the curve feature according to the correction value; the feature configuration parameters are used for updating model parameters of a preset clustering model, and the preset model parameters are kmeans mean clustering models.

Step S1422, extracting the current model parameters of the preset clustering model, and calculating the current clustering accuracy of the preset clustering model according to the correlation coefficient between the current model parameters and the feature configuration parameters; and when the current clustering accuracy is lower than a set accuracy, updating the preset clustering model by using the characteristic configuration parameters.

Step S1423, determining a clustering feature number based on the updated preset clustering model, and judging whether the vector dimension number of the word vector is equal to the clustering feature number; if not, adjusting the vector dimension degree of the word vector based on the feature configuration parameters so that the vector dimension degree of the word vector is equal to the clustering feature number.

Step S1424, inputting the adjusted word vector into the updated preset clustering model and obtaining at least two clustering sets output by the updated preset clustering model; and aiming at each cluster set, calculating cosine distances between each word vector in the cluster set and other word vectors, screening the cosine distances according to the clustering weight of each word vector in the cluster set, and calculating the mean value of the cosine distances obtained by screening to obtain the feature difference.

Thus, by performing the steps S1421 to S1424, the clustering accuracy and the feature difference accuracy can be ensured.

In one possible embodiment, in order to ensure reliable verification of a plurality of tag information, on the basis of step S120, what is described in the following step S121 and step S122 may be further included.

Step S121, acquiring a modification instruction for modifying the set identification degree and the set similarity rate.

And step S122, modifying the set identification degree and the set similarity according to the modification instruction.

Therefore, the set identification degree and the set similarity can be flexibly modified according to different service scenes, and reliable verification of the information of the plurality of labels is ensured.

In an alternative embodiment, the step S130 may specifically include the following steps S131 to S135, which are described in detail below, of generating a time sequence continuity curve of the message in the current time period according to the confidence of the message in each target tag information, the information identification of each target tag information, and the tag similarity between the target tag information, and extracting the keyword queue of the message in the current time period.

Step S131, the confidence degrees of the message under each target label information are sequenced from large to small to obtain a confidence degree sequence, and a plurality of time sequence nodes of the message in the current time period are determined according to the confidence degree sequence and the information identification degree of each target label information.

Step S132, fitting the plurality of time sequence nodes based on the label similarity rate between the target label information to obtain an initial curve, and smoothing the initial curve by adopting the median in the confidence coefficient sequence to obtain the time sequence continuity curve.

Step S133, respectively determining the associated message fields of the information identification degree of each target tag information in the message, and generating the word segmentation identifier of the message according to the field characteristics of each associated message field.

Step S134, splitting the message by using the word segmentation identification to obtain a plurality of message sequences.

Step S135, determining an extraction logic based on the tag similarity between the target tag information, extracting the keyword in each packet sequence through the extraction logic, and combining the keywords according to the generation time information of the corresponding packet sequence to obtain the keyword queue.

When the contents described in the above steps S131 to S135 are applied, the smoothness of the time series continuity curve and the integrity of the keyword queue can be ensured.

Based on the same inventive concept, please refer to fig. 2 in combination, a functional block diagram of aninformation processing apparatus 200 based on live webcasting and online electronic commerce delivery is provided, where theinformation processing apparatus 200 is applied to the cloud server, and specifically includes the following functional blocks:

the taginformation determining module 210 is configured to determine, for each intelligent terminal, information identification degrees of a plurality of tag information to be verified, which are used for determining time sequence continuity of the interactive information and extracting keywords of the interactive information, and tag similarity rates between different tag information, based on the extracted first information list and second information list for recording the interactive information of the intelligent terminal;

a taginformation verification module 220, configured to verify the plurality of tag information based on the determined information identification degrees of the plurality of tag information and tag similarity rates between different tag information, and count at least three pieces of target tag information that pass verification; the information identification degree of the target label information is greater than the set identification degree, and the label similarity rate between the target label information is smaller than the set similarity rate;

amessage processing module 230, configured to obtain a message of the intelligent terminal in a current time period determined based on the operation log of the cloud server, generate a time sequence continuity curve of the message in the current time period according to a confidence of the message in each target tag information, an information identification degree of each target tag information, and a tag similarity rate between the target tag information, and extract a keyword queue of the message in the current time period;

the intelligentterminal detection module 240 is configured to determine whether the intelligent terminal carries a trojan program based on the time sequence continuity curve and the keyword sequence; if so, setting an access interception mechanism for intercepting the intelligent terminal according to the terminal configuration information of the intelligent terminal.

For the description of the functional modules, please refer to the description of the method shown in fig. 1, which is not described herein again.

Based on the same inventive concept, as shown in fig. 3, an architecture diagram of aninformation processing system 300 based on live webcasting and online e-commerce delivery is provided. Theinformation processing system 300 may include acloud server 310 and a plurality ofsmart terminals 320 in communication with thecloud server 310, thecloud server 310 being configured to:

for each intelligent terminal, determining information identification degrees of a plurality of label information to be verified, which are used for determining the time sequence continuity of the interactive information and extracting keywords of the interactive information, and label similarity rates among different label information based on a first information list and a second information list which are extracted and used for recording the interactive information of the intelligent terminal;

verifying the plurality of label information based on the determined information identification degrees of the plurality of label information and the label similarity rates among different label information, and counting at least three target label information passing the verification; the information identification degree of the target label information is greater than the set identification degree, and the label similarity rate between the target label information is smaller than the set similarity rate;

acquiring a message of the intelligent terminal in a current time period determined based on an operation log of the cloud server, generating a time sequence continuity curve of the message in the current time period according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate among the target label information, and extracting a keyword queue of the message in the current time period;

judging whether the intelligent terminal carries a Trojan program or not based on the time sequence continuity curve and the keyword sequence; if so, setting an access interception mechanism for intercepting the intelligent terminal according to the terminal configuration information of the intelligent terminal.

Further, the determining, by thecloud server 310, whether the intelligent terminal carries the trojan program based on the time sequence continuity curve and the keyword sequence specifically includes:

obtaining a critical curve segment in the time sequence continuity curve, wherein the critical curve segment comprises a curve segment with a curve slope larger than a critical point of a set slope; equally dividing the critical curve segment into at least two sub-line segments; calculating the curve slope change coefficient included in each sub-line segment of the at least two sub-line segments in parallel; mapping each calculated curve slope change coefficient to a preset value interval and calculating the weighted sum of the mapping values of each curve slope change coefficient in the preset value interval;

when the weighted sum is larger than a target value, extracting a word vector of each keyword in the keyword sequence; clustering the extracted multiple word vectors based on curve features corresponding to the time sequence continuity curves to obtain at least two clustering sets; calculating the feature difference degree between the word vectors in each cluster set;

when the difference value of each feature difference degree and the set difference degree is smaller than a preset threshold value, selecting a cluster set corresponding to the maximum feature difference degree as a detection set, determining a word vector matrix corresponding to the detection set, and searching whether a target word vector matrix with the matrix similarity between the word vector matrix and the word vector matrix larger than the set similarity exists or not from a preset database; if yes, judging that the intelligent terminal carries the Trojan horse program; and matrix elements in the target word vector matrix are elements corresponding to abnormal keywords.

On the basis of the above, please refer to fig. 4, which provides acloud server 310, which includes aprocessor 311 and amemory 312, which are in communication with each other, wherein theprocessor 311 retrieves a computer program from thememory 312 and runs the computer program, so as to implement the method shown in fig. 1.

Furthermore, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when running, implements the method as shown in fig. 1.

In summary, based on the method, the apparatus, the system, and the cloud server, the information identification degrees of a plurality of tag information and the tag similarity rates between different tag information are determined according to the first information list and the second information list.

And secondly, verifying the plurality of label information based on the determined information identification degrees of the plurality of label information and the label similarity rate among different label information, and counting at least three target label information passing the verification.

And then, acquiring the message of the intelligent terminal in the current time period, generating a time sequence continuity curve of the message according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate among the target label information, and extracting a keyword queue of the message.

And finally, setting an access interception mechanism when the intelligent terminal is judged to carry the Trojan horse program based on the time sequence continuity curve and the keyword sequence.

Therefore, whether the intelligent terminal carries the Trojan program or not can be timely and accurately detected by carrying out time sequence continuity detection and natural language analysis on the interactive information of the intelligent terminal, so that the intelligent terminal carrying the Trojan program is accessed and intercepted, and the information safety of an e-commerce platform and other intelligent terminals is ensured.

It is to be understood that the present invention is not limited to what has been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the invention is limited only by the appended claims.

Claims (7)

1. An information processing method based on network live broadcast and online e-commerce delivery is characterized by being applied to a cloud server communicating with a plurality of intelligent terminals, and at least comprising the following steps:

for each intelligent terminal, determining information identification degrees of a plurality of label information to be verified, which are used for determining the time sequence continuity of the interactive information and extracting keywords of the interactive information, and label similarity rates among different label information based on a first information list and a second information list which are extracted and used for recording the interactive information of the intelligent terminal;

verifying the plurality of label information based on the determined information identification degrees of the plurality of label information and the label similarity rates among different label information, and counting at least three target label information passing the verification; the information identification degree of the target label information is greater than the set identification degree, and the label similarity rate between the target label information is smaller than the set similarity rate;

acquiring a message of the intelligent terminal in a current time period determined based on an operation log of the cloud server, generating a time sequence continuity curve of the message in the current time period according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate among the target label information, and extracting a keyword queue of the message in the current time period;

judging whether the intelligent terminal carries a Trojan program or not based on the time sequence continuity curve and the keyword sequence; if so, setting an access interception mechanism for intercepting the intelligent terminal according to the terminal configuration information of the intelligent terminal;

generating a time sequence continuity curve of the message in the current time period and extracting a keyword queue of the message in the current time period according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate among the target label information, and the method comprises the following steps: sequencing the confidence degrees of the message under each target label information according to a descending order to obtain a confidence degree sequence, and determining a plurality of time sequence nodes of the message in the current time period according to the confidence degree sequence and the information identification degree of each target label information; fitting the plurality of time sequence nodes based on the label similarity rate between the target label information to obtain an initial curve, and smoothing the initial curve by adopting a median in the confidence coefficient sequence to obtain a time sequence continuity curve; respectively determining the associated message fields of the information identification degree of each target label information in the message, and generating word segmentation identifiers of the message according to the field characteristics of each associated message field; splitting the message by adopting the word segmentation identification to obtain a plurality of message sequences; and determining extraction logic based on the label similarity rate between the target label information, extracting the key words in each message sequence through the extraction logic, and combining the key words according to the generation time information of the corresponding message sequence to obtain the key word queue.

2. The method of claim 1, wherein determining whether the intelligent terminal carries a Trojan program based on the time sequence continuity curve and the keyword sequence comprises:

obtaining a critical curve segment in the time sequence continuity curve, wherein the critical curve segment comprises a curve segment with a curve slope larger than a critical point of a set slope; equally dividing the critical curve segment into at least two sub-line segments; calculating the curve slope change coefficient included in each sub-line segment of the at least two sub-line segments in parallel; mapping each calculated curve slope change coefficient to a preset value interval and calculating the weighted sum of the mapping values of each curve slope change coefficient in the preset value interval;

when the weighted sum is larger than a target value, extracting a word vector of each keyword in the keyword sequence; clustering the extracted multiple word vectors based on curve features corresponding to the time sequence continuity curves to obtain at least two clustering sets; calculating the feature difference degree between the word vectors in each cluster set;

when the difference value of each feature difference degree and the set difference degree is smaller than a preset threshold value, selecting a cluster set corresponding to the maximum feature difference degree as a detection set, determining a word vector matrix corresponding to the detection set, and searching whether a target word vector matrix with the matrix similarity between the word vector matrix and the word vector matrix larger than the set similarity exists or not from a preset database; if yes, judging that the intelligent terminal carries the Trojan horse program; and matrix elements in the target word vector matrix are elements corresponding to abnormal keywords.

3. The method according to claim 2, wherein the extracted multiple word vectors are clustered based on curve features corresponding to the time series continuity curve to obtain at least two cluster sets, and a feature difference degree between the word vectors in each cluster set is calculated, further comprising:

determining the characteristic dimension quantity of the curve characteristic, normalizing the position information of each characteristic value in the characteristic value queue corresponding to the curve characteristic to obtain a relative position coefficient corresponding to each characteristic value, correcting each characteristic value by using the relative position coefficient to obtain a correction value corresponding to each characteristic value, and calculating a characteristic configuration parameter corresponding to the curve characteristic according to the correction value; the characteristic configuration parameters are used for updating model parameters of a preset clustering model, and the preset model parameters are kmeans mean clustering models;

extracting current model parameters of the preset clustering model, and calculating the current clustering accuracy of the preset clustering model according to a correlation coefficient between the current model parameters and the feature configuration parameters; when the current clustering accuracy is lower than a set accuracy, updating the preset clustering model by using the characteristic configuration parameters;

determining a clustering feature number based on the updated preset clustering model, and judging whether the vector dimension number of the word vector is equal to the clustering feature number; if not, adjusting the vector dimension number of the word vector based on the feature configuration parameters so as to enable the vector dimension number of the word vector to be equal to the clustering feature number;

inputting the adjusted word vectors into the updated preset clustering model and obtaining at least two clustering sets output by the updated preset clustering model; and aiming at each cluster set, calculating cosine distances between each word vector in the cluster set and other word vectors, screening the cosine distances according to the clustering weight of each word vector in the cluster set, and calculating the mean value of the cosine distances obtained by screening to obtain the feature difference.

4. The method according to any one of claims 1 to 3, wherein, based on the extracted first information list and second information list for recording the interaction information of the intelligent terminal, determining the information identification degrees of a plurality of label information to be verified for determining the time sequence continuity of the interaction information and for extracting the keywords of the interaction information, and the label similarity ratio between different label information, further comprises:

listing a first information set used for representing communication records between the intelligent terminal and the cloud server in the first information list according to a time sequence, listing a second information set used for representing communication records between the intelligent terminal and other intelligent terminals in the second information list according to the time sequence, generating a first information flow track of the first information list based on the first information set, and generating a second information flow track of the second information list based on the second information set; the number of first track nodes in the first information flow track is different from the number of second track nodes in the second information flow track, and the first track nodes and the second track nodes have different node information dimensions;

acquiring a time sequence continuous track for representing the interactive information and target information of an information subject of the interactive information, which are represented by a node information dimension corresponding to one first track node in the first information flow track; determining a second track node with the largest node information dimension in the second information flow track as a target track node, and packaging the target information into the target track node according to a time sequence difference value between the first information flow track and the second information flow track so as to obtain configuration information corresponding to the target information in the target track node; determining a time sequence consistency weight between the first information list and the second information list according to the matching degree between the configuration information and the target information;

performing time sequence correction on the first information stream track and the second information stream track by adopting the time sequence consistency weight, and splicing the corrected first information stream track and the corrected second information stream track to obtain a third information stream track; extracting node description information corresponding to each third track node in the third information flow track and classifying the node description information corresponding to each third track node to obtain a first information group and a second information group;

determining a plurality of pieces of first label information to be verified for determining the time sequence continuity of the interactive information based on the node description information in the first information group, determining a plurality of pieces of second label information to be verified for determining the keyword for extracting the interactive information based on the node description information in the second information group, and screening out a plurality of pieces of third label information to be verified for determining the time sequence continuity of the interactive information and extracting the keyword for the interactive information according to the first label information and the second label information; mapping the label index value corresponding to each piece of third label information to a preset numerical value list to obtain the information identification degree corresponding to each piece of third label information; and calculating cosine distances between information vectors of every two pieces of third label information, and weighting the calculated cosine distances to obtain label similarity rates among the plurality of pieces of third label information.

5. The method according to claim 4, wherein the classifying the node description information corresponding to each third trace node to obtain a first information packet and a second information packet specifically comprises:

dividing a track curve corresponding to the third information stream track into a plurality of curve segments according to a preset time interval, calculating a curve slope for representing the track stability of each curve segment, and generating a track distribution list corresponding to the third information stream track according to the curve slope; classifying the node description information corresponding to each third track node for the first time based on the information association degree of the list information in the track distribution list to obtain a first initial group and a second initial group;

when it is determined that an information compatibility identifier exists between the node description information under the first initial grouping and the node description information under the second initial grouping, adjusting at least one node description information under the first initial grouping to be under the second initial grouping based on the determined information compatibility between the node description information under the first initial grouping and the node description information under the second initial grouping;

calculating a first information configuration coefficient of the node description information under the first initial grouping after the adjustment and a second information configuration coefficient of the node description information under the second initial grouping after the adjustment are completed, and determining a coefficient difference value between the first information configuration coefficient and the second information configuration coefficient; judging whether the coefficient difference is larger than a set difference or not; when the coefficient difference is larger than the set difference, determining a first initial packet after adjustment as the first information packet and determining a second initial packet after adjustment as the second information packet; and when the coefficient difference is smaller than or equal to the set difference, continuing to adjust at least one piece of node description information under the first initial grouping to the second initial grouping, and returning to the step of calculating a first information configuration coefficient of the adjusted node description information under the first initial grouping and a second information configuration coefficient of the adjusted node description information under the second initial grouping.

6. The method of claim 1, further comprising:

acquiring a modification instruction for modifying the set identification degree and the set similarity rate;

and modifying the set identification degree and the set similarity according to the modification instruction.

7. An information processing system based on network live broadcast and online electronic commerce delivery is characterized by comprising a cloud server and a plurality of intelligent terminals communicated with the cloud server; wherein the cloud server is configured to:

for each intelligent terminal, determining information identification degrees of a plurality of label information to be verified, which are used for determining the time sequence continuity of the interactive information and extracting keywords of the interactive information, and label similarity rates among different label information based on a first information list and a second information list which are extracted and used for recording the interactive information of the intelligent terminal;

verifying the plurality of label information based on the determined information identification degrees of the plurality of label information and the label similarity rates among different label information, and counting at least three target label information passing the verification; the information identification degree of the target label information is greater than the set identification degree, and the label similarity rate between the target label information is smaller than the set similarity rate;

acquiring a message of the intelligent terminal in a current time period determined based on an operation log of the cloud server, generating a time sequence continuity curve of the message in the current time period according to the confidence of the message in each target label information, the information identification degree of each target label information and the label similarity rate among the target label information, and extracting a keyword queue of the message in the current time period;

judging whether the intelligent terminal carries a Trojan program or not based on the time sequence continuity curve and the keyword sequence; if so, setting an access interception mechanism for intercepting the intelligent terminal according to the terminal configuration information of the intelligent terminal.

Images