Detailed Description
The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
It should be noted that, the information and data related to the user in the embodiments of the present disclosure are information and data authorized by the user or fully authorized by the related parties, and the processes of collecting, storing, using, processing, transmitting, providing, disclosing and applying the related data all comply with relevant laws and regulations and standards, take necessary security measures, do not violate the public welcome, and provide corresponding operation entries for the user or the related parties to select authorization or rejection.
It should also be noted that in the embodiments of the present disclosure, some existing solutions in the industry such as software, components, models, etc. may be mentioned, and they should be considered as exemplary, only for illustrating the feasibility of implementing the technical solution of the present disclosure, but not meant to imply that the applicant has or must not use the solution.
The technical terms related to the invention are as follows:
An APP Application;
API Application Programming Interface application program interfaces;
SDK Software Development Kit software development kits;
WIFI WIRELESS FIDELITY networks;
UI User Interface user interfaces;
URL Uniform Resource Locator uniform resource locator.
Due to the impact of network performance on the user experience, feedback will be very direct to the operation of the service. Moreover, the mobile network has inherent weak network problems, DNS problems, connectivity performance, etc. that cannot be compared to conventional fixed networks. Therefore, the invention provides a method for optimizing the mobile terminal network and the friendly reminding popup window display.
As shown in fig. 1, the embodiment of the present disclosure provides a fault-reporting popup window page guiding method based on weak network detection.
The method is particularly applied to the side of the client. In particular implementations, the method may include the following:
S101, detecting the current network state of a user in real time, and generating a network detection result according to a preset network detection rule;
S102, if the network detection result is network abnormality, acquiring current webpage browsing information of a user, wherein the current webpage browsing information comprises a current user tag, a current user behavior habit and a current user behavior track;
S103, generating a recommended guide result according to a network detection result, user webpage browsing information and a pre-established network monitoring guide strategy and/or a popup guide prediction model, wherein the network monitoring guide strategy is realized by carrying out aggregation analysis based on historical user webpage browsing information, and the popup guide prediction model is realized by utilizing a training set and a testing set which are obtained based on the historical user webpage browsing information, and is realized by carrying out random initialization based on hidden layer parameters and solving the weight of the hidden layer output to an output layer;
And S104, outputting guide information according to the recommended guide result to prompt a user to execute corresponding operation based on the guide information.
The network detection result may specifically include timeout of network response, disconnection of network, whether the network belongs to a slow request, whether the network is in an illegal network segment or WIFI segment, whether DNS/domain name resolution is accessed normally, and the like.
Specifically, in order to meet the requirements of the service product line, the client may generate a weak network detection framework (network detection policy) in advance, and based on the preset network detection policy, a corresponding network detection result may be generated according to the network to be detected.
The weak network detection framework can be dynamically injected into a local client through CCB (Church Community Builder) API, specifically, the weak network detection framework is dynamically injected into a host APP through CCB (Church Community Builder) API, and asynchronous online update configuration table, update initialization strategy and method are carried out to provide the environment for the next network detection.
In specific implementation, the network detection policy may be implemented by classifying and defining each stage of the network, classifying and defining a total amount of states of each network node layer, extracting characteristic values of the states to standard and standardize each node, where the network detection policy includes, but is not limited to, network detection rules in the following 5:
1. and detecting that the network response time exceeds the set time and belongs to time-out, and recording the state 0.
The set duration may be, for example, 15s, and may be specifically set according to needs, which is not limited to the present invention.
2. Network disconnection-record state 1 for no access network relationship.
3. And (3) network slow/weak network, namely monitoring the slow request of which the network response time exceeds the set time for 3s, and recording the state 2.
The set time length can be set according to the needs, and can be determined according to the time-out time length which can be tolerated by the user at the big data statistics place.
4. Network early warning, namely detecting an illegal network segment or a public unsafe WIFI segment, and recording a state 3.
The unsafe WIFI section is determined according to a preset relationship table corresponding to the unsafe WIFI section, and the relationship table can be updated in real time.
5. DNS/domain name resolution-detecting if DNS/domain name resolution is normally accessed.
The network monitoring guiding policy comprises popup window guiding, network abnormal page prompting, guiding a user to set or related network detection and check, managing background resource bit newly added application configuration support and the like.
To meet the business product line requirements, the client may pre-create a popup management framework (i.e., network listening guideline policy) and/or a popup guideline prediction model. The popup management framework and the popup guiding prediction model can be dynamically injected into a local client through CCB (Church Community Builder) APIs, specifically, CCB (Church Community Builder) APIs are dynamically injected into a host APP, an asynchronous online update configuration table, an update initialization strategy and a method are performed, and the environment for monitoring and calling of the next network is provided. Network listening guidance policies include, but are not limited to, the following:
1. And dynamically detecting the timeout of the network and friendly prompting popup window, namely timeout of network link and operation guidance.
2. And (5) network disconnection reminding, namely, network link disconnection reminding and operation guidance.
3. Cellular/WiFi handoff cues that cellular traffic operation guidelines are being used.
4. Weak network prompt popup window, namely network link bad operation guidance.
5. And prompting abnormal domain name resolution, namely prompting abnormal domain name resolution and guiding operation.
In specific implementation, the configuration strategy is needed to be carried out on the indexes and the nodes of each stage of the network so as to meet the flow guidance and popup error reminding required by each functional scene.
The call related to network detection and guidance needs to be dynamically injected into codes in advance, and the call is business logic which needs to be processed after the APP is started, so that the call has the actions of inputting and outputting the internal and external API methods, including associated strategies, error code prompt documents and the like.
The network API classification configuration table may be subdivided into:
1. And (3) registering APiconfig.json of the full-quantity API registration table, registering APIs used by the full quantity of each platform channel, and carrying out classification marking.
The full API registry is an API registry for network popup scheduling used in the multi-terminal multi-platform, and is uniformly responsible for popup and scheduling use by the SDK in the basic API. According to the requirement, a privately-owned custom interface can be realized to realize custom error reporting popup.
The full API registry is an API of the registered network call layer, and is common to multiple platforms. There is a clear description for each API and a scenario correspondence usage description, including rule authority restrictions, etc.
The description of the data structure field and parameters in the field of apicon fig. json is as follows:
Unique identification (id): "0001", which is a class defined unique identifier used to ensure the uniqueness of the API interface in the system.
Description (des): "sweep API" is used to briefly describe the function or use of the interface.
Platform "iOS/android" means that the API interface is suitable for use with mobile platforms, including android and iOS operating systems.
Channel type (type): "ccbmobile/ccblife", indicates a specific use channel of the API, such as an application platform for mobile banking or institutional life.
Android class name (android_ classname): "com.ccb.home.appscreen.ccbbridge", which is the full name of the class implementing the API function on the Android platform.
IOS class name (ios_ classname): "CCBSceneBridge", which is the class name that implements the API function on the iOS platform.
API name (APIname): "STARTSCENEAPI", which is the name of the API, is used to identify its function.
Parameter (params): "ke1=value & ke2=value 2", which means the parameters and their corresponding values required when the API is called, support delivery in the form of key-value pairs.
Rights rule (rule): "tag: _00028| branchid:440000000", this field defines the rights rules required to initiate a popup. The rules in the example indicate that the customer must meet the conditions of a particular tag and a segment number in order to use the relevant function.
The data structure is used for describing a specific API interface, and through the fields, the data structure provides a comprehensive definition, and covers basic information, applicable platform, class name, parameter configuration and authority requirements of the API so as to facilitate development and use.
Fig. 2 is a scheduling flow chart of the APIs corresponding to the full API registry, and as shown in fig. 2, the flow scheduling process includes:
S201, API call.
And S202, packaging the CCBH5 Bridge level, and distributing tasks to a scheduling layer for indicating how to pop windows and prompt.
And S203, distributing to a CCBH5 Operation scheduling policy layer.
S204, judging whether the user has authority, whether the user has a corresponding calling method, and whether the aeronaut has respective implementation methods.
S205, if yes, triggering the corresponding method call and the callback, and if no, entering S206.
S206, ending.
2. Rule authority configuration table ruleconfig. json, details of usage for types, rules, parameters, etc. that require specific association of configuration APIs.
The rule fields of a certain scenario are as follows:
"params":"key1=value1",
"des" scans the API, is dedicated for mobile banking,
"rule":"",
The above fields are registered and described for part of a particular API, with different rules being used for different scenarios.
According to the configured strategy, the network state value of the current user can be clarified through searching, calculating and monitoring of each network node, and user guidance is given, so that a complete user network popup window guidance flow closed loop is achieved, and better user interaction experience is achieved.
Taking network acquisition as an example for illustration:
The JSON data records specific information of the network request, and is convenient for analysis and fault detection. Each request is identified by a unique event ID and contains the start and end times (in milliseconds) of the request, thereby calculating the delay of the request. The scene ID defines the particular environment, such as a network or advertisement, in which the request occurs.
The result of the request is indicated by the response code, a successful request returns the status code 200, and a failed request is identified by 404. The error description field provides detailed information of network layer anomalies, and the status code field further indicates whether the request was successful (0 indicates success, 1 indicates failure). In addition, the response status code and message field provide supplemental information regarding the outcome of the request.
The URL of the request and its parameters contain the specific content of the request. In the trace section, the behavior of the request in different processing nodes (e.g., HTTP, DNS, and TCP) is recorded, including the processing time of each stage, the corresponding error information, and its success status. The structured records enable performance monitoring and problem analysis of network requests to be clearer, help teams identify specific links of delay or errors, and therefore effective optimization and adjustment are carried out.
The above examples of network acquisition are mainly aimed at the indexes and descriptions related to network performance monitoring and acquisition, specifically, indexes such as a certain index, such as a certain model, a certain time point, a certain error reporting information, a time consuming time, and the like, so as to comprehensively judge network performance and error reporting direction.
In some embodiments, as shown in fig. 3, the foregoing detecting the current network state of the user in real time, generating the network detection result according to the preset network detection rule, and when implemented, may include the following:
s301, detecting the network state of the current network according to a preset network state strategy;
s302, acquiring a network detection strategy which is dynamically injected to a local client through a CCB API;
s303, inquiring a corresponding network detection rule from a network detection strategy according to the current network state;
S304, acquiring network detection parameters of the current network;
And S305, detecting and matching the network state according to the network detection rule and the network detection parameter to obtain a corresponding network detection result.
The network state may specifically include a network normal and a network abnormal. In a normal network state, the invention generally does not need operation guidance. Network anomalies may include network timeouts, network outages, slow/weak networks, network early warning, detecting if DNS/domain name resolution is normally accessed, etc.
It should be noted that, in S301, detecting the current network state may be regarded as a preliminary detection method of network detection, and the purpose is to query and verify the network detection rule based on the preliminary detection result.
The preliminary detection can be performed according to preset network state policies, wherein the preset network state policies include longer network connection time (longer than a set time), incapable network connection, slower network speed (lower than a preset minimum network speed), prompt that the network may have risks, incapable normal access of DNS/domain name resolution (DNS server failure, domain name expiration, DNS setting errors, and the like), and the like.
As described above, the network detection policy stores a plurality of network detection rules, and based on an abnormal situation in the current network state, the client may query one or more network detection rules that may correspond to the network detection policy.
Specifically, the network detection parameters of the current network may include, for example, a network connection duration exceeding 30 seconds, a network speed below 50k/s, and so on.
In some embodiments, as shown in fig. 4, the detecting and matching the network state according to the network detection rule and the network detection parameter to obtain a corresponding network detection result may include the following when implemented:
s401, matching network detection rules with network detection parameters to network states;
S402, comparing the network detection rule obtained by matching with the network detection parameter, and obtaining a corresponding network detection result according to the comparison result;
S403, acquiring an error code pre-allocated to the network detection result.
In S401, the network detection rules in the network detection policy need to be matched with the network detection parameters respectively to obtain network detection rules related to the network detection parameters, for example, the network transmission speed is low, for example, one of the network detection parameters is 5k/S, so that the network slow/weak network problem may exist, and the network detection rules- "network slow/weak network: monitoring network response duration exceeds the slow request of the set duration for 3S" can be matched, and the state 2 "is recorded. It should be noted that the network detection parameters may include a plurality of network detection rules that are matched to the network detection parameters.
And comparing the network detection rule obtained by matching with the network detection parameter, and obtaining a corresponding network detection result according to whether the comparison result meets the preset condition. For example, the current network connection duration is 30 seconds, and the preset set duration is 15s, and the network detection result is that the network response is overtime.
Specifically, each network detection result may be assigned an error code to be used as an error code hint when performing operation guidance.
Specifically, in order to meet the line requirements of the service product, the preset network state policy and the network detection policy are generated in advance. The preset network state strategy and network detection strategy can be dynamically injected into the host APP of the local client through CCBAPI, and asynchronous online updating of the configuration table and updating of the initialization strategy and method are performed.
In some embodiments, as shown in fig. 5, the steps of the pre-created popup prediction model may include the following when implemented:
S501, acquiring historical user webpage browsing information, wherein the historical user webpage browsing information comprises user labels, historical user behavior habits and historical user behavior tracks;
specifically, the client acquires historical user webpage browsing information of a plurality of users from the server as a model training sample, wherein the historical data comprises user tags, historical user behavior habits and historical user behavior tracks, and the historical data is used for performing model training to obtain a popup guide prediction model.
S502, preprocessing web browsing information of a historical user to obtain a training set and a testing set;
Specifically, the client preprocesses the web browsing information of the historical user, including data cleaning, data standardization, data complement, data division and the like, removes low-quality data, and respectively obtains a corresponding high-quality training set and a corresponding high-quality testing set so as to utilize the high-quality training set and the high-quality testing set to perform model training to generate a popup window guide prediction model.
And S503, performing model training by using a training set and a testing set, and generating a popup window guide prediction model based on random initialization of hidden layer parameters and solving of weights output from a hidden layer to an output layer.
Specifically, the client side respectively carries out model training by using a training set and a testing set which are obtained by processing the webpage browsing information of the historical user, and obtains a popup guide prediction model. The popup guidance prediction model is obtained based on neural network training. The client randomly initializes the weight and bias from the input layer to the hidden layer, then carries out nonlinear transformation on the input data through the hidden layer to obtain a hidden layer output matrix, and finally solves the weight output from the hidden layer to the output layer by utilizing the hidden layer output matrix and the target output of the training data. The input quantity of the popup guide prediction model comprises a user tag, a historical user behavior habit and a historical user behavior track, and the output quantity is a recommended guide result.
The user labels specifically comprise factors such as gender, age, roles, regions and the like, the historical user behavior habits comprise factors such as user preference, browsing duration, user behavior data points and the like, the historical user behavior tracks comprise data such as user clicking and browsing tracks, the user labels and the historical user behavior habits are static attributes of users, and the historical user behavior tracks are dynamic attributes of the users.
In one embodiment, the client selects data of different age groups (for example, 10 years is an age group) in the historical user web browsing information, each age group selects 2000 groups of data, and the popup guide prediction model of each age group is obtained through training.
In one embodiment, the client performs model training based on a Single hidden Layer feedforward neural network (Single-Layer Feedforward Neural Network, SLFN), which is composed of an input Layer, a hidden Layer, and an output Layer, including but not limited to a Perceptron (Perceptron), a Multi-Layer Perceptron (MLP), a radial basis function network (Radial Basis Function Network, RBFN), a Self-Organizing Map (SOM), and an Extreme learning machine (Extreme LEARNING MACHINE, ELM).
In an embodiment, the client trains based on an Extreme learning machine (Extreme LEARNING MACHINE, ELM) to obtain a popup prediction model, and the structure of the popup prediction model is shown in fig. 11, wherein the input weight and the hidden layer threshold are randomly generated, and the output layer weight matrix is obtained by one-step analytic calculation independently of training sample data.
For an observation sample (xi,ti), where the input vector xi＝[xi1,xi2,xi3,…xin]T∈Rn, (i=1, 2,3,., N), N being the number of observation samples, N is the dimension number of the sample input vector, and the output vector ti＝[ti1,ti2,…tim]T∈Rm, m is the dimension of the output sample, namely the node number of the output layer of the ELM model. Let the number of hidden layer nodes of the model be l, the excitation function be g (·), then the model of ELM is:
Where βi＝[βi1,βi2,…,βim]T is the weight from the ith node of the hidden layer to the output layer of the model, wi＝[wi1,wi2,…,win]T is the weights of the ith nodes of the input layer and the hidden layer of the model, and bi represents the threshold value of the ith node of the hidden layer. The ELM model output values may be zero error fit samples, i.e.:
I.e. the presence of βi,wi and bi satisfies:
Can be abbreviated as:
Hβ=T (4)
Wherein:
Where H is referred to as the hidden layer output matrix. In the model training stage, the input weight and the bias value of the feedforward neural network are randomly set, and the ELM learning training problem is converted into the least square norm problem for solving the output weight matrix beta by calculating the output matrix H, namely:
wherein H+ is the generalized inverse of matrix H.
The client side respectively carries out model training by utilizing training sets and test sets obtained by processing the webpage browsing information of the historical users of different age groups to obtain popup guide prediction models of different age groups. The input quantity of the popup guide prediction model of different age groups comprises user labels, historical user behavior habits and historical user behavior tracks, and the output quantity is a recommended guide result.
In the training embodiment, the user is divided into age groups, which is not limited by the application, and the age groups may not be divided during specific implementation.
In some embodiments, a recommended guidance result may be generated according to a network detection result, user web browsing information, and a popup guidance prediction model, and guidance information may be output according to the recommended guidance result to prompt a user to perform a corresponding operation based on the guidance information, where the method includes the following steps:
1. format for outputting guidance information
The recommended guidance results are converted into user-friendly information formats such as popup notifications, message reminders for mobile applications, card patterns in web pages or email notifications, and the like.
2. Operation prompt design
Each prompt contains an explicit Action Call (Call to Action), such as "try reload page", "check network settings", "recommend you to access this link", etc., encouraging the user to take specific actions.
Preferably, a user feedback mechanism can be integrated, so that the user is allowed to evaluate the recommended guide information, and the recommendation model is further optimized and the user experience is improved.
The network monitoring guiding policy is implemented by aggregation analysis based on the web browsing information of the historical user, and specifically can comprise the following contents:
user tag, user history behavior habit and user history behavior track
1. User behavior analysis
And analyzing browsing habits and behavior patterns of the users through a statistical analysis method, for example, grouping the users according to similar browsing behaviors through cluster analysis, and identifying different types of users. The access patterns of the user in different time periods can be analyzed through time series analysis, and the high-frequency access period and abnormal behaviors can be identified.
2. Problem identification
Based on the browsing history of the user, potential network problems are identified. For example, for network timeout problems, it may be analyzed whether a user frequently experiences delays in accessing certain particular websites. For the off-network case, it may be checked whether the user has no access to the network for a certain period of time. For domain name errors, the analysis may be performed by the user accessing 404 the error page that appears in the record.
3. Generating guideline policies
And generating personalized operation guidance according to the historical behaviors and labels of the user aiming at the identified problems. For example, for a particular user, a timeout may occur when it is detected that he frequently visits a certain website, the user may be advised to choose a more stable network connection or to change DNS settings, and for users who frequently experience a network disconnection problem, they may be advised to check router settings or contact a network service provider.
Preferably, after implementing the guideline policy, user feedback is collected, validity of the guideline is evaluated, and the analysis model and the guideline policy are continuously optimized and adjusted according to the feedback data. In addition, the data model can be updated in real time by continuously tracking the browsing behavior and the network condition of the user, so that the accuracy and pertinence of the aggregation analysis are improved, and a dynamic network monitoring guiding strategy is realized.
In some embodiments, as shown in fig. 6, the step of creating the network listening guidance policy in advance may include the following steps when implemented:
S601, acquiring historical user webpage browsing information, wherein the historical user webpage browsing information comprises user tags, user historical behavior habits and user historical behavior tracks;
s602, determining the preference degree of a user according to a user tag, a user historical behavior habit and a user historical behavior track based on a collaborative filtering recommendation algorithm;
And S603, generating a network monitoring guide strategy according to the determined preference degree.
The specific obtaining method of the user tag can include collecting the user tag in the processes of user registration, information input and the like. Additionally, user tags may be dynamically adjusted based on the user's online behavior (e.g., clicks, purchases, etc.) to maintain accuracy and relevance of information.
The user history behavior habit may be the behavior of the user in a period of time, including browsing duration, page access frequency, clicking behavior, etc. Various behavioral habits of the user may be recorded in real time using website analysis tools (e.g., google analysis) or custom tracking codes. The historical behavior habits of the user may be stored in a time-series manner to form a user behavior log for subsequent use.
The user history behavior trace may be a browsing path of the user within a specific period of time, including pages accessed, time stamps, event types, etc. Recording of the user's historical behavior trace may use event driven methods, recording relevant information in real time and uploading to a back-end database when the user accesses a page, clicks on a link. The user historical behavior tracks are regularly tidied and cleaned so as to optimize the storage and searching efficiency, and a certain historical record is reserved for model training and analysis.
The collaborative filtering recommendation algorithm which can be based on when determining the preference degree of the user comprises a collaborative filtering algorithm based on the user, a collaborative filtering algorithm based on the object, and the like. In the collaborative filtering algorithm based on the user, the content can be recommended by comparing the similarity between different users, for example, viewing the browsing records of the user A and the user B, if the users have the same interest content, the content which is interested by the user B but not seen by the user A can be recommended, and in the collaborative filtering algorithm based on the object, the object which is loved by the user in the past can be analyzed, and then other objects similar to the objects are recommended.
And then, calculating the similarity between the users and the articles by using a similarity calculation method (such as cosine similarity, euclidean distance, pearson correlation coefficient and the like), predicting potential scores of the users on the unvisited content based on the behavior records of the known users, and generating preference scores by weighted average to realize score presetting. And then, the preference degree classification is needed, the calculated scores are standardized, and a threshold value is set to distinguish high-preference content, medium-preference content and low-preference content so as to better generate a recommendation strategy.
The network monitoring guiding strategy is realized by carrying out aggregation analysis based on the webpage browsing information of the historical user, and when the network monitoring guiding strategy is generated according to the determined preference degree, the personalized network monitoring guiding strategy can be formulated according to the labels of the user, the historical behavior habits and the analysis results of the behavior tracks. The recommendation behavior in a specific situation is explicitly defined, e.g. if the user preference is above a certain threshold, an update of the content of this type is recommended. Preferably, the present application also allows for the design of adjustable parameters that allow the policy to be dynamically adjusted based on real-time user feedback (e.g., user behavior changes, preference shifts).
Based on the obtained network monitoring guiding policy, a recommended guiding result can be generated according to the network detection result, the user webpage browsing information and the network monitoring guiding policy, and the method can be realized in the following manner:
1. Recommendation guideline generation process
And data integration, namely acquiring network state data and online behaviors of the current user, and combining the previously acquired user browsing history information and the formulated guiding strategy.
And the effect evaluation mechanism is used for performing effect analysis on the generated guiding results by using real-time data input, such as calculating the click rate of the actual recommended content and the deep participation degree of the user so as to evaluate the effectiveness and rationality of the guiding strategy.
2. Generation and optimization of recommendation results
And the personalized recommendation result is that a recommendation content list aiming at each user is generated based on the generated network monitoring guide strategy, wherein the recommendation content list comprises various forms such as text information, pictures and videos, so that the attraction is improved.
And the continuous optimization mechanism is used for periodically analyzing historical data and user feedback, optimizing a recommendation algorithm and a strategy and ensuring that a recommendation result is always consistent with user requirements.
Based on the obtained recommended guidance result, the guidance information may be output to prompt the user to execute the corresponding operation based on the guidance information, which may be specifically implemented in the following manner:
1. Display form of guide information
The mode of variously expressing the recommended content comprises popup window prompt, page navigation bar recommendation, mail notification and the like, so that the user is ensured to see the recommended information at proper time.
2. User operation guidance
And a simple and clear operation guide is provided, the operation guide is connected with the recommended content, the meaning and the effect of the recommendation are ensured to be understood by the user, and the user is encouraged to make corresponding operations (such as clicking a link, purchasing recommended goods and the like).
3. User feedback collection and looping
After each recommendation, a user feedback form is added, so that the user is allowed to evaluate (like, dislike and suggestion) the recommended content, and the data is fed back to the recommendation algorithm to perform TM optimization so as to continuously improve the personalized experience of the user.
The network monitoring guiding policy is implemented by performing aggregation analysis based on the web browsing information of the historical user, and in some embodiments, as shown in fig. 7, the step of creating the network monitoring guiding policy in advance may include the following steps when implemented:
S701, acquiring historical user webpage browsing information, wherein the historical user webpage browsing information comprises user labels, user historical behavior habits and user historical behavior tracks;
s702, obtaining interest tags of users according to user tags, user historical behavior habits and user historical behavior tracks based on a content recommendation algorithm;
and S703, obtaining a network monitoring guide strategy of user preference according to the interest tag.
The obtaining of the user tag, the user history behavior habit and the user history behavior track are described in detail in the above embodiment corresponding to fig. 6, and are not described herein.
In the formation process of the interest labels, the labels, behavior habits and historical tracks of the users can be integrated, and the content recommendation algorithm is used for generating the interest labels of the users. For example, users frequently browse web pages of the "science and technology", "travel" categories for which relevant interest tags (e.g., "science and technology lovers", "travel explorers") are to be generated. The interest labels can be verified and optimized regularly through A/B tests or user feedback, and the accuracy and instantaneity of the generated interest labels are ensured.
Specifically, the network monitoring guiding strategy for obtaining the user preference according to the interest tag can be realized by the following method:
1. formulation of network monitoring policy
And according to the generated interest labels, a specific network monitoring strategy is formulated to monitor content variation and information update highly related to the user interest. Policy elements may include the type of content to be monitored (e.g., news, product updates, social media trends), the particular website or platform of interest, the frequency of information updates (e.g., real-time monitoring, timing queries), etc.
2. Application of personalized policies
Personalized implementations are made, for example, if the user tag shows that it is interested in "healthy life", the policy will focus on monitoring information on relevant web pages for healthy products, fitness activities, and diet advice. In addition, the network monitoring strategy can be dynamically adjusted according to the change of the user behavior and the update of the interest labels, so that the latest requirements of the users are always met.
And according to the network detection result, the user webpage browsing information and the network monitoring guide strategy, firstly integrating and processing data when generating a recommended guide result, and then generating the recommended guide result. When data integration is carried out, the acquired real-time data (such as price change, user evaluation, new product release and the like), the historical user webpage browsing information and the network monitoring guide strategy are integrated into a comprehensive data set. When data processing is performed, a big data processing framework (such as APACHE SPARK or Hadoop) is used for cleaning, converting and analyzing data to extract valuable information. When the recommendation guiding result is generated, a recommendation algorithm (such as collaborative filtering or content filtering) can be used for generating the personalized recommendation guiding result according to the integrated data. Preferably, a real-time feedback mechanism can be introduced, and the effectiveness of the recommendation result is evaluated according to the click rate of the user and the interaction condition, so that the recommendation algorithm is continuously optimized.
The form of outputting the guide information can be various in the process of prompting the user to execute the corresponding operation based on the guide information according to the recommended guide result, including but not limited to push notification, in-station message, email, social media sharing and the like, so as to ensure timeliness and effectiveness of the information. Through designing a user-friendly interface, the recommended guide information is presented in an intuitively understandable manner.
When the user is prompted to execute corresponding operation based on the guiding information, operation prompt design can be performed, and the output information contains clear operation prompts, such as 'click view details', 'immediate purchase', and the like, so that the user is encouraged to perform further interaction according to the recommendation result.
Preferably, in the output of the recommendation result, a user feedback option can be added to evaluate the satisfaction of the recommended content. By collecting user rating data, the recommendation algorithm and the guideline policy are optimized. And combining user feedback with behavior data to form virtuous circle as the basis for generating interest labels and formulating network monitoring strategies in the next round, so as to continuously improve user experience.
The generating of the recommended guidance result according to the network detection result, the user webpage browsing information and the pre-created network monitoring guidance policy specifically may include matching the corresponding network monitoring guidance policy according to the network detection result and the user webpage browsing information as the recommended guidance result. In an embodiment, according to the network detection result and the user web browsing information matching the corresponding network monitoring guiding policy, the similarity between the network detection result and the user web browsing information and the web browsing information of each user can be calculated, and the similarity can be cosine similarity, euclidean distance, pearson correlation coefficient and the like.
In the embodiments corresponding to fig. 5 to fig. 7, the generation of the recommended guidance result according to the network detection result, the user web browsing information and the network monitoring guidance policy, and the generation of the recommended guidance result according to the network detection result, the user web browsing information and the popup guidance prediction model are described. In the application, the recommended guide result can be generated according to the network detection result, the user webpage browsing information, the network monitoring guide strategy and the popup guide prediction model.
The recommendation guide result can be generated based on the network monitoring guide strategy, and the recommendation guide result can be generated based on the popup guide prediction model. In addition, better recommended guide results can be obtained based on the recommended guide results generated by the network monitoring guide strategy and the recommended guide results generated by the popup window guide prediction model. For example, better recommended guidance results of the two are output to the user based on the preference of the user.
In some embodiments, as shown in fig. 8, the generating a recommended guidance result according to the network detection result, the user web browsing information, and the network monitoring guidance policy and/or the popup guidance prediction model may include the following when implemented:
S801, generating a first recommended guide result according to the network detection result, the user webpage browsing information and a pre-established network monitoring guide strategy;
S802, generating a second recommended guide result according to the network detection result, the user webpage browsing information and a pre-established popup guide prediction model;
S803, obtaining preference settings of the user, and respectively calculating the relatedness of the first recommended guide result and the second recommended guide result to the preference settings, wherein the recommended guide result with high relatedness is used as a final recommended guide result.
The first recommended guiding result may refer to the embodiment corresponding to fig. 6 and fig. 7, and the second recommended guiding result may refer to the embodiment corresponding to fig. 5, which is not described in detail.
In S803, the user preference settings may be acquired in various ways, options are provided to the user at the first run of the system, user registration or user setting interface, and the user preference settings are collected in the form of a questionnaire, a selection question or a slider, etc. The preference settings may include, but are not limited to, content types (e.g., news, entertainment, science, sports, etc.), language preferences, price ranges (for e-commerce recommendations), brand preferences, and update frequency (e.g., daily, weekly, etc.), etc. The preference settings are stored in a user profile database and associated with a unique identifier of the user (e.g., user ID) for subsequent retrieval and use.
In specific implementation, a relevance calculating technology (such as cosine similarity, a jaccard similarity coefficient or a euclidean distance) may be used to calculate a relevance between each first recommended guidance result and the user preference setting, and a relevance between each second recommended guidance result and the user preference setting. For example, for each recommended content, features (e.g., categories, keywords) are extracted and compared to user preferences, and the results of the calculation are saved in the form of scores, with higher scores indicating more relevant.
After the correlation between the first recommended guidance result and the user preference setting and the correlation between the second recommended guidance result and the user preference setting, the correlation scores can be integrated, the correlation scores of the first recommended guidance result and the second recommended guidance result are sorted, and the scores of the corresponding recommended results are compared. A weighted average approach may be used to weight different recommended sources (e.g., the first recommendation result is weighted higher) in order to emphasize more reliable recommended sources. When determining the recommendation guiding result with high correlation, a threshold can be set, and the recommendation content with the correlation exceeding a certain threshold is screened out, so that the final recommendation guiding result is used as the final recommendation guiding result, and the final recommendation result not only needs to meet the preference setting of the user, but also needs to be matched with the interests of the user. And displaying the screened final recommendation guiding result to the user. At the moment, the system can tell the user about the detected matching condition of the user preference and the recommended content together so as to enhance the user experience, and meanwhile, the transparency of the system and the trust of the user are improved.
Fig. 9 is a schematic diagram of UI popup window guidance in the embodiment of the present disclosure, fig. 10 is a schematic diagram of weak network guidance in the embodiment of the present disclosure, where the network guidance and popup window guidance may include classifying conditions such as network timeout, disconnection, weak network, DNS, link, domain name resolution, etc., designing corresponding error codes and popup window guidance prompts, network anomaly page prompts, guiding a user to set up by himself or check a related network, managing background resource newly added application configuration support, etc., which is not limited in the present disclosure.
From the above, based on the error report popup window page guiding method provided by the embodiment of the specification, when a user uses a network, the user can detect the current network state in real time, generate a network detection result according to a preset network detection rule, acquire the current user tag, the current user behavior habit and the current user behavior track if the network detection result is abnormal, generate a recommended guiding result according to the network detection result, the user web browsing information and a pre-established network monitoring guiding strategy and/or popup window guiding prediction model, and output guiding information according to the recommended guiding result to prompt the user to execute corresponding operation based on the guiding information. Therefore, aggregation analysis and popup window guiding can be carried out on function inlets of different users, and effective scheduling of popup window guiding of a weak network is achieved.
The embodiment of the present disclosure provides an electronic device, and is shown in fig. 11. The electronic device includes a network communication port 1101, a processor 1102, and a memory 1103, where the foregoing structures are connected by an internal cable, so that each structure may perform specific data interaction.
The network communication port 1101 may be specifically configured to receive current user web browsing information, where the current user web browsing information includes a current user tag, a current user behavior habit, and a current user behavior track.
The processor 1102 is specifically configured to generate a network detection result according to a preset network detection rule, obtain current web browsing information of a user if the network detection result is a network anomaly, generate a recommended guidance result according to the network detection result, the user web browsing information, and a pre-created network monitoring guidance policy and/or a popup guidance prediction model, where the network monitoring guidance policy is implemented by performing an aggregation analysis based on historical user web browsing information, the popup guidance prediction model is implemented by using a training set and a test set obtained based on the historical user web browsing information, and is implemented by randomly initializing based on hidden layer parameters and solving a hidden layer output to an output layer weight, and output guidance information according to the recommended guidance result to prompt the user to perform a corresponding operation based on the guidance information.
The memory 1103 may be specifically configured to store relevant data such as a corresponding instruction program and a target processing rule.
Based on the method, the related structural performance of the server can be effectively utilized, the data processing speed of the electronic equipment is improved, and the data processing of the market management bill is efficiently realized.
In this embodiment, the network communication port 1101 may be a virtual port that binds with different communication protocols, so that different data may be sent or received. For example, the network communication port may be a port responsible for performing web data communication, a port responsible for performing FTP data communication, or a port responsible for performing mail data communication. The network communication port may also be an entity's communication interface or a communication chip. For example, it may be a wireless mobile network communication chip such as GSM, CDMA, etc., it may also be a Wifi chip, it may also be a bluetooth chip.
In this embodiment, the processor 1102 may be implemented in any suitable manner. For example, a processor may take the form of, for example, a microprocessor or processor, and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, application SPECIFIC INTEGRATED Circuits (ASICs), programmable logic controllers, and embedded microcontrollers, among others. The description is not intended to be limiting.
In this embodiment, the memory 1103 may include multiple levels, and in a digital system, the memory may be any memory as long as binary data can be stored, in an integrated circuit, a circuit with a storage function without a physical form, such as a RAM, a FIFO, etc., and in a system, a storage device with a physical form, such as a memory bank, a TF card, etc., may also be a memory.
The embodiment of the specification also provides a computer readable storage medium of the error reporting popup page guiding method, wherein the computer readable storage medium stores computer program instructions, when the computer program instructions are executed, the computer program instructions are realized by generating network detection results according to preset network detection rules, acquiring current webpage browsing information of a user if the network detection results are abnormal, generating recommended guiding results according to the network detection results, the user webpage browsing information and a pre-established network monitoring guiding strategy and/or popup guiding prediction model, the network monitoring guiding strategy is realized by carrying out aggregation analysis based on historical user webpage browsing information, the popup guiding prediction model is realized by utilizing training sets and testing sets obtained based on historical user webpage browsing information, random initialization based on hidden layer parameters and solving of hidden layer output to output layer weights, and outputting guiding information according to the recommended guiding results to prompt a user to execute corresponding operations based on guiding information.
In the present embodiment, the storage medium includes, but is not limited to, a random access Memory (Random Access Memory, RAM), a Read-Only Memory (ROM), a Cache (Cache), a hard disk (HARD DISK DRIVE, HDD), or a Memory Card (Memory Card). The memory may be used to store computer program instructions. The network communication unit may be an interface for performing network connection communication, which is set in accordance with a standard prescribed by a communication protocol.
In this embodiment, the functions and effects of the program instructions stored in the computer readable storage medium may be explained in comparison with other embodiments, and are not described herein.
The embodiment of the specification also provides a computer program product, which at least comprises a computer program, wherein the computer program is executed by a processor and realizes the following method steps of generating a network detection result according to a preset network detection rule, acquiring current webpage browsing information of a user if the network detection result is abnormal, generating a recommended guidance result according to the network detection result, the user webpage browsing information and a pre-established network monitoring guidance strategy and/or popup guidance prediction model, wherein the network monitoring guidance strategy is realized by carrying out aggregation analysis based on historical user webpage browsing information, the popup guidance prediction model is realized by utilizing a training set and a testing set which are obtained based on the historical user webpage browsing information, and is realized by randomly initializing hidden layer parameters and solving hidden layer output to an output layer weight, and outputting guidance information according to the recommended guidance result so as to prompt the user to execute corresponding operation based on the guidance information.
Referring to fig. 12, the embodiment of the present disclosure further provides a fault-reporting popup window page guiding device based on weak network detection, where the device may specifically include the following structural units:
The network detection result comprises a network response timeout, whether a network is disconnected, whether the network belongs to a slow request, whether the network is in an illegal network segment or a WIFI segment, and whether DNS/domain name resolution is normally accessed;
The information acquisition unit 1202 is used for detecting the current network state of a user in real time and generating a network detection result according to a preset network detection rule, wherein the network detection result comprises overtime network response, disconnection of a network, slow request of the network, illegal network section or WIFI section of the network and normal access of DNS/domain name resolution;
The guiding result generating unit 1203 is configured to generate a recommended guiding result according to a network detection result, user web browsing information, and a pre-created network monitoring guiding policy and/or popup guiding prediction model, where the network monitoring guiding policy is implemented by performing aggregation analysis based on historical user web browsing information, and the popup guiding prediction model is implemented by using a training set and a testing set obtained based on the historical user web browsing information, based on random initialization of hidden layer parameters, and solving the weight output to the output layer by the hidden layer;
and an information output unit 1204 for outputting guide information according to the recommended guide result to prompt the user to perform a corresponding operation based on the guide information.
In one embodiment, the detection result generating unit comprises a state detection module, a strategy acquisition module, a rule query module, a parameter acquisition module and a detection matching module.
The state detection module is used for detecting the network state of the current network according to a preset network state strategy;
The strategy acquisition module is used for acquiring the network detection strategy, and the network detection strategy is dynamically injected to the local client through the CCB API;
The rule query module is used for querying a corresponding network detection rule from the network detection strategy according to the network state;
the parameter acquisition module is used for acquiring network detection parameters of the current network;
And the detection matching module is used for detecting and matching the network state by the network detection rule and the network detection parameter to obtain a corresponding network detection result.
In one embodiment, the detection matching module comprises a state matching sub-module, a comparison sub-module and an error code distribution module.
The state matching sub-module is used for matching the network state by the network detection rule and the network detection parameter;
the comparison submodule is used for comparing the network detection rule obtained by matching with the network detection parameter, and obtaining a corresponding network detection result according to the comparison result;
the error code distribution module is used for obtaining the error codes which are distributed in advance by the network detection result.
In one embodiment, the error-reporting popup page guiding device further comprises a model creating unit, wherein the model creating unit comprises a historical information acquisition module, a preprocessing module and a model generating module.
The historical information acquisition module is used for acquiring historical user webpage browsing information, wherein the historical user webpage browsing information comprises user tags, historical user behavior habits and historical user behavior tracks;
The preprocessing module is used for preprocessing the historical user webpage browsing information to obtain a training set and a testing set;
the model generation module is used for carrying out model training by utilizing the training set and the testing set, and generating a popup window guiding prediction model based on random initialization of hidden layer parameters and solving of weights output to an output layer by the hidden layer.
In one embodiment, the error-reporting popup page guiding device further comprises a history information acquisition unit, a user preference determination unit and a strategy generation unit.
The history information acquisition unit is used for acquiring the history user webpage browsing information, wherein the history user webpage browsing information comprises user tags, user history behavior habits and user history behavior tracks;
The user preference determining unit is used for determining the preference degree of the user according to the user label, the user historical behavior habit and the user historical behavior track based on the collaborative filtering recommendation algorithm;
and the strategy generating unit is used for generating the network monitoring guide strategy according to the determined preference degree.
In one embodiment, the error-reporting popup page guiding device further comprises a history information acquisition unit, a label generation unit and a label generation unit.
The history information acquisition unit is used for acquiring the history user webpage browsing information, wherein the history user webpage browsing information comprises user tags, user history behavior habits and user history behavior tracks;
the label generating unit is used for obtaining interest labels of users according to the characteristic description events of the user service products based on the content recommendation algorithm;
and the strategy determining unit is used for obtaining the network monitoring guide strategy of the user preference according to the interest tag.
In an embodiment, the guiding result generating unit 1203 is specifically configured to match a corresponding network monitoring guiding policy according to the network detection result and the user web browsing information, and use the network monitoring guiding policy as a recommended guiding result.
In one embodiment, the guiding result generating unit 1203 includes a first result generating module, a second result generating module, and a guiding result generating module.
The first result generation module is used for generating a first recommended guide result according to the network detection result, the user webpage browsing information and a pre-established network monitoring guide strategy;
the second result generation module is used for generating a second recommended guide result according to the network detection result, the user webpage browsing information and a pre-established popup guide prediction model;
The guide result generation module is used for acquiring preference settings of a user, and calculating correlation degrees of the first recommended guide result and the second recommended guide result with the preference settings respectively, wherein the recommended guide result with high correlation degrees is used as a final recommended guide result.
From the above, based on the error report popup window page guiding device provided by the embodiment of the specification, when a user uses a network, the user can detect the current network state in real time, generate a network detection result according to a preset network detection rule, acquire the current user tag, the current user behavior habit and the current user behavior track if the network detection result is abnormal, generate a recommended guiding result according to the network detection result, the user web browsing information and a pre-established network monitoring guiding strategy and/or popup window guiding prediction model, and output guiding information according to the recommended guiding result to prompt the user to execute corresponding operation based on the guiding information. Therefore, aggregation analysis and popup window guiding can be carried out on function inlets of different users, and effective scheduling of popup window guiding of a weak network is achieved.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In the description of the present specification, reference to the terms "one embodiment," "one particular embodiment," "some embodiments," "for example," "an example," "a particular example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the application, and is not meant to limit the scope of the application, but to limit the application to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the application are intended to be included within the scope of the application.