Disclosure of Invention
Based on this, it is necessary to provide a chat group creation method, apparatus, computer device and storage medium for intelligently matching appropriate users according to real-time chat log information of the users to aggregate users having the same interests in the same chat group, aiming at the drawbacks of the existing chat group creation method.
In order to achieve the above object, the present invention provides a chat group creation method, including:
when the user matching function is detected to be triggered, acquiring a target chat text to be analyzed and processed;
analyzing and predicting the target chat text based on a pre-trained prediction model, and acquiring user information to be matched based on an obtained analysis and prediction result;
when an input selection instruction is received, matching the matched user selected in the user information to be matched with a target user corresponding to the target chat text, and establishing a chat group.
Optionally, before the target chat text for analysis processing is acquired when the user matching function is detected to be triggered, the method further includes:
when receiving instruction information for user matching sent by a first terminal, determining whether a second terminal in communication connection with the first terminal exists or not;
when determining that a second terminal in communication connection with the first terminal exists, sending the instruction information to the second terminal;
triggering the user matching function when a determining instruction which is sent by the second terminal and allows user matching is received within a preset time length;
And when the second sent determining instruction allowing user matching is not received within the preset time, regarding the instruction information as an invalid instruction and sending corresponding prompt information to the first terminal.
Optionally, when receiving instruction information for user matching sent by the first terminal, determining whether a second terminal in communication connection with the first terminal exists, and then the method further includes:
and triggering the user matching function when the second terminal which is in communication connection with the first terminal is determined to be absent.
Optionally, when the user matching function is detected to be triggered, acquiring the target chat text for analysis processing includes:
when a user matching instruction is received, acquiring time information corresponding to the user matching instruction;
and acquiring the chat text within a preset time period based on the time information as a target chat text.
Optionally, the analyzing and predicting the target chat text based on the pre-trained prediction model, and acquiring the user information to be matched based on the obtained analyzing and predicting result, includes:
preprocessing the target chat text to obtain a preprocessed target chat text;
Inputting the preprocessed target chat text into a preset convolutional neural network to obtain a target label corresponding to the target chat text output by the convolutional neural network;
and carrying out query matching in a preset record list based on the target label to obtain the user information to be matched, wherein the preset record list records the corresponding relation between the user information and the label information.
Optionally, the obtaining the user information to be matched by performing query matching in a preset record list based on the target tag includes:
acquiring a preset record list, and reading first tag information corresponding to user information contained in the preset record list;
calculating a similarity value between the first tag information and the target tag;
comparing the similarity value with a preset threshold value;
and when the similarity value is greater than or equal to the preset threshold value, determining that the user information corresponding to the first tag information is the user information to be matched.
Optionally, the method further comprises:
when a model training instruction is received, receiving the input chat text to be classified and the classification label;
training the chat text to be classified and the classification label as input of a prediction model to be trained;
And when the fact that the prediction model to be trained starts to converge is detected, determining that the training of the prediction model to be trained is completed, and obtaining the trained prediction model.
In addition, in order to achieve the above object, the present invention also provides a chat group creation apparatus including:
the text acquisition module is used for acquiring a target chat text for analysis processing when the user matching function is detected to be triggered;
the user determining module is used for carrying out analysis and prediction on the target chat text based on a pre-trained prediction model and obtaining user information to be matched based on an obtained analysis and prediction result;
and the group creation module is used for matching the selected matched user with the target user corresponding to the target chat text when receiving the input selection instruction, and creating a chat group.
Optionally, the chat group creation device further includes:
the first determining module is used for determining whether a second terminal which is in communication connection with the first terminal exists or not when receiving instruction information which is sent by the first terminal and is used for user matching;
the instruction sending module is used for sending the instruction information to a second terminal which is in communication connection with the first terminal when the second terminal is determined to exist;
The function triggering module is used for triggering the user matching function when receiving a determining instruction which is sent by the second terminal and allows user matching in a preset time length; and the second terminal is used for sending the instruction information to the first terminal as an invalid instruction when the second sent determining instruction which allows the user to be matched is not received within the preset time.
Optionally, the instruction sending module is specifically further configured to: and triggering the user matching function when the second terminal which is in communication connection with the first terminal is determined to be absent.
Optionally, the text obtaining module includes:
the time acquisition unit is used for acquiring time information corresponding to the user matching instruction when the user matching instruction is received;
and the text acquisition unit is used for acquiring the chat text in the preset time period based on the time information as a target chat text.
Optionally, the user determining module includes:
the preprocessing unit is used for preprocessing the target chat text to obtain a preprocessed target chat text;
the text input unit is used for inputting the preprocessed target chat text into a preset convolutional neural network to obtain a target label corresponding to the target chat text output by the convolutional neural network;
And the user query unit is used for carrying out query matching in a preset record list based on the target label to obtain the user information to be matched, and the preset record list records the corresponding relation between the user information and the label information.
Optionally, the user query unit includes:
the label acquisition subunit is used for acquiring a preset record list and reading first label information corresponding to user information contained in the preset record list;
a similarity calculating subunit, configured to calculate a similarity value between the first tag information and the target tag;
a similarity comparison subunit, configured to compare the similarity value with a preset threshold;
and the user determination subunit is used for determining that the user information corresponding to the first tag information is the user information to be matched when the similarity value is greater than or equal to the preset threshold value.
Optionally, the chat group creation device further includes:
the information receiving module is used for receiving the input chat text to be classified and the classification label when receiving the model training instruction;
the model training module is used for training the chat text to be classified and the classification label as input of a prediction model to be trained;
And the detection completion module is used for determining that the training of the prediction model to be trained is completed when the convergence of the prediction model to be trained is detected, so as to obtain the trained prediction model.
In addition, to achieve the above object, the present invention further provides a computer device, including a memory and a processor, where the memory stores computer readable instructions, where the computer readable instructions, when executed by the processor, cause the processor to perform the steps of the chat group creation method described above.
In addition, to achieve the above object, the present invention also provides a storage medium storing computer readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the chat group creation method described above.
When the user matching function is detected to trigger starting, the method, the device, the computer equipment and the storage medium for creating the chat group firstly acquire the target chat text for analysis and processing, then analyze and predict the acquired target chat text by utilizing a pre-trained prediction model, further acquire corresponding user information to be matched according to the acquired analysis and prediction result, and finally match the selected skin Pi A user with the target user corresponding to the target chat text when receiving the input selection instruction so as to form the created chat group. The method and the device realize intelligent matching of proper users according to the real-time chat record information of the users, and aggregate the users with the same interests in the same chat group, so that the real-time accuracy of chat group creation is improved.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless expressly stated otherwise, as understood by those skilled in the art. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
Referring to fig. 1, fig. 1 is a flow chart illustrating a chat group creation method in an embodiment.
In one embodiment, the chat group creation method includes:
step S10, when the user matching function is detected to be triggered, a target chat text to be analyzed and processed is obtained;
the user matching function is a function of matching users with the same chat characteristics to form a chat group. The triggering of the user matching function is realized by the corresponding operation of the user, for example, the key on the terminal is touched, and the virtual key or the physical key can be used. Since the generation of the chat log requires at least two users to be in a chat state, the triggering of the user matching function requires all agreements of the users in the same chat group as the current user, that is, the triggering of the user matching function is triggered under the condition that all users agree.
When receiving the user matching instruction, the chat group creation device acquires the target chat text which needs to be analyzed and processed. In practical application, when matching chat users, users with proper chat needs to be matched with current users, so that analysis and determination are needed to be performed on current chat information of users to know chat characteristics corresponding to the users, such as chat subjects or topics, and then users with the same chat characteristics can be matched with the current users, and a chat topic ring is formed.
Further, when the target chat text is acquired, the method specifically includes:
step a, when a user matching instruction is received, acquiring time information corresponding to the user matching instruction;
and b, acquiring chat texts in a preset time period based on the time information to serve as target chat texts.
When receiving the user matching instruction, the chat group creation device acquires the time information of the user matching instruction, acquires a target chat text according to the acquired time information and a preset text acquisition rule, and particularly acquires the chat text in a preset time period as the target chat text.
The target chat text is an object of analysis processing by the chat group creation device, but in practical application, the chat record text of the user is huge, and only the recorded chat text exists, so that a more suitable chat text needs to be acquired in the history chat text as the target chat text, and a specific manner of acquiring the target chat text can be acquired according to time information or according to occurrence time of keyword information, and is not limited in this embodiment.
Step S20, analyzing and predicting a target chat text based on a pre-trained prediction model, and acquiring user information to be matched based on an obtained analysis and prediction result;
when the chat group creating device obtains the target chat text subjected to analysis processing, the target chat text needs to be analyzed and predicted, and further user information for matching, which can be matched, can be obtained according to analysis and prediction results, wherein the obtained user information of the user to be matched can comprise identification information of the user, can also comprise personal information of the user, and can specifically perform corresponding setting on the type of the information for matching.
Specifically, when the chat group creation device acquires the target chat text, a pre-trained prediction model is first acquired, and then the target chat text is analyzed and processed by using the obtained prediction model. In practical application, each user allowed to match chat users performs feature analysis processing according to the corresponding target chat text, so that each user has a corresponding feature tag, and users with the same feature tag have a certain matching association relationship, namely users to be matched, which are matched by opposite parties.
In this embodiment, when the target text is analyzed by using the prediction model, the target chat text is mainly analyzed by using a natural language processing (Natural Language Processing, NLP) technology, so as to obtain a corresponding analysis and prediction result. In practical application, a pre-trained prediction model (Text-CNN model) is utilized to analyze and process a target chat Text, specifically, the target chat Text is used as input of the prediction model, and then an analysis and prediction result of the target chat Text is output, wherein the analysis and prediction result can specifically include feature labels.
In addition, in order to accurately perform analysis processing on the Text-CNN model, training of the model is required in advance. When training the model, taking the chat text as input of a corpus as the model, setting corresponding characteristic labels, training the model, and analyzing and processing the chat text according to the trained prediction model.
And step S30, when an input selection instruction is received, matching the matched user selected in the user information to be matched with a target user corresponding to the target chat text, and establishing a chat group.
When receiving an input selection instruction, the chat group creation device firstly determines the currently selected matching user according to the selection instruction, and then establishes a matching relationship between the matching user and a target user corresponding to the target chat text so as to establish and form a corresponding chat group. In practical application, after the chat text of the user is analyzed and processed, the user which can be matched is obtained, at the moment, the user to be matched which can be matched can be directly matched with the current user to form a corresponding chat group, the user can also perform corresponding selection, and a proper and proper number of users are selected to be matched to form the chat group.
In this embodiment, when the user matching function is detected to trigger and start when the chat group is created, firstly, a target chat text to be analyzed is obtained, then, the obtained target chat text is analyzed and predicted by using a pre-trained prediction model, further, corresponding user information to be matched is obtained according to the obtained analysis and prediction result, and finally, when an input selection instruction is received, the selected skin Pi A user is matched with the target user corresponding to the target chat text, so as to form the created chat group. The method and the device realize intelligent matching of proper users according to the real-time chat record information of the users, and aggregate the users with the same interests in the same chat group, so that the real-time accuracy of chat group creation is improved.
Further, referring to fig. 2, fig. 2 is a flowchart illustrating steps triggered by the user matching function in one embodiment.
Specifically, in step S10, when it is detected that the user matching function is triggered, before the target chat text for performing the analysis processing is acquired, the method further includes:
step S40, when receiving instruction information for user matching sent by the first terminal, determining whether a second terminal in communication connection with the first terminal exists;
step S50, when determining that a second terminal in communication connection with the first terminal exists, sending the instruction information to the second terminal;
step S60, when a determination instruction which is sent by the second terminal and allows user matching is received within a preset time period, triggering a user matching function;
and step S70, when a second sent determination instruction allowing user matching is not received within a preset time period, the instruction information is regarded as an invalid instruction, and corresponding prompt information is sent to the first terminal.
The terminal is a device used by the chat user, including but not limited to a mobile phone, and a certain association relationship exists between the terminal and the chat user, for example, the chat user uses the terminal to log in an account. The "first" and "second" of the "first terminal" and the "second terminal" are used only for distinguishing terminals, and in practical application, the first terminal and the second terminal may be interchanged. In addition, step S60 and step S70 are parallel schemes, and only one of them will exist.
When receiving instruction information for user matching sent by a first terminal, the chat group creation device firstly determines whether a second terminal for communication connection with the first terminal exists, when determining that the second terminal for communication connection with the first terminal exists, the chat group creation device sends the received instruction information for user matching sent by the first terminal to the second terminal, further receives feedback information of the second terminal based on the instruction information, when receiving a determination instruction for allowing user matching sent by the first terminal within a preset time period, the chat group creation device triggers a corresponding user matching function, and when not receiving the determination instruction for allowing chat user matching sent by the second terminal within the preset time period, the chat group creation device regards the instruction information sent by the first terminal as an invalid instruction and sends corresponding prompt information to the first terminal.
In practical application, when users match chat users, one user wants to find a person or group with the same chat topic, or one chat group wants to find more persons or groups with the same chat topic or interest, and for different situations, the processing modes of the chat group creation device are different, so when the chat group creation device receives instruction information for matching chat users sent by the first terminal, it is first determined whether there are second terminals in communication connection with the first terminal, where the number of the second terminals is not limited, and the number of the second terminals can be one or multiple. When a second terminal which is in communication connection with the first terminal exists, the chat group creation device sends instruction information sent by the first terminal to the second terminal, so that the second terminal determines whether user matching can be performed or not, and when the second terminal determines that user matching can be performed, the chat group creation device starts a user matching function of the chat group creation device to perform user matching.
Because the number of the second terminals is not limited, only one user is required to determine whether to perform user matching when the number of the second terminals is one, and not all users want to perform user matching when the number of the second terminals is more than two, so that when the number of the second terminals is more than or equal to two, the chat group creation device can determine whether to start the user matching function according to the received feedback information of the second terminals. For example, when the number of the second terminals is ten, when the first terminal sends out instruction information for user matching, the ten second terminals all receive the instruction information forwarded by the chat creation device, then the ten second terminals feed back the received instruction information, wherein the specific feedback comprises agreement and disagreement, and finally, whether user matching is performed is determined according to feedback information sent out by the ten second terminals. For the case that the number of the second terminals is smaller, it may be determined that user matching can be performed only when all the second terminals agree, and when the number of the second users is too large, it may be determined according to the ratio conditions of different decisions in the feedback information, for example, when the feedback information is that the agreed ratio is greater than or equal to 80%, it is determined that user matching can be performed, and obviously, the ratio is customizable.
In addition, there is a second terminal that is communicatively connected to the first terminal, and there is a case where no communication connection is made to the first terminal, that is, a case where an independent user wants to chat with a user or group having the same interest topic, at this time, when the first terminal sends a user matching instruction, and there is no second terminal that is communicatively connected to the first terminal, the chat group creation device will directly start the user matching function.
Further, referring to fig. 3, fig. 3 is a flowchart illustrating a step of analyzing and predicting a target chat text to obtain user information to be matched in an embodiment.
Specifically, step S20, performing analysis and prediction on the target chat text based on the pre-trained prediction model, and obtaining the user information to be matched based on the obtained analysis and prediction result, includes:
step S21, preprocessing the target chat text to obtain a preprocessed target chat text;
step S22, inputting the preprocessed target chat text into a preset convolutional neural network to obtain a target label corresponding to the target chat text output by the convolutional neural network;
step S23, query matching is carried out in a preset record list based on the target label, user information to be matched is obtained, and the corresponding relation between the user information and the label information is recorded in the preset record list.
After receiving the obtained target chat text, the chat group creating device analyzes and predicts the target chat text by utilizing a pre-trained prediction model, specifically, firstly, pre-processes the received target chat text to obtain a pre-processed target chat text, then, taking the pre-processed target chat text as input of a preset convolutional neural network to obtain a target label corresponding to the target chat text output by the convolutional neural network, and finally, carrying out query matching in a preset record list according to the obtained target label to obtain user information to be matched.
In practical application, when analyzing and predicting a target chat text, when acquiring the target chat text, the chat group creating device firstly performs preprocessing on the target chat text, mainly including word segmentation, stop word removal and the like, and then performs vectorization on the preprocessed target chat text, and by preprocessing the target chat text, invalid information in the target chat text is removed, so that the obtained text information is more accurate.
When analysis and prediction are carried out, the text is mainly analyzed and processed by utilizing an NLP technology, specifically, the training of a prediction model is carried out in advance based on a convolutional neural network, so that the target chat text is analyzed and predicted by utilizing the trained prediction model. Different classification association rules of the chat text and the labels exist in the pre-trained prediction model, after the target chat text subjected to pretreatment is input into the convolutional neural network, the target labels corresponding to the target chat text can be directly input, and the number of the target labels can be one or a plurality of target labels. The prediction model needs to be trained before being used, specifically, when a high-quality and large-data-volume supervised corpus is input, including chat text and classification labels, then a set of classification rules, namely the prediction model, are automatically trained in a machine learning mode, and then the input chat text to be marked is marked and classified when the input chat text to be marked is received.
Further, referring to fig. 4, fig. 4 is a flowchart illustrating a step of obtaining user information to be matched in one embodiment.
Specifically, step S23, performing query matching in a preset record list based on the target tag, to obtain user information to be matched, includes:
step S231, a preset record list is obtained, and first tag information corresponding to user information contained in the preset record list is read;
step S232, calculating a similarity value between the first tag information and the target tag;
step S233, comparing the similarity value with a preset threshold;
in step S234, when the similarity value is greater than or equal to the preset threshold, it is determined that the user information corresponding to the first tag information is the user information to be matched.
The preset record list is recorded with relevant information of users or groups which are allowed to be matched with others, and the relevant information comprises label information corresponding to the users or groups.
After the chat group creation device predicts the target chat text to obtain a corresponding target label, determining user information to be matched according to the target label, specifically, after the target label is obtained, obtaining a preset record list, reading first label information corresponding to the user information contained in the preset record list, calculating a similarity value of the first label information and the target label, further obtaining corresponding user information to be matched according to the similarity value, specifically comparing the obtained similarity value with a preset threshold, and determining that the user information corresponding to the first label information is the user information to be matched when the obtained similarity value is greater than or equal to the preset threshold.
In practical application, a plurality of users and group information are recorded in a preset record list, different users and groups are respectively associated with own tag information, when a target tag is obtained, similarity values between the target tag and tag information corresponding to all users or groups in the preset record list are calculated, and whether the corresponding user information can be used as user information to be matched is determined according to the size of the similarity values. For example, the number of the preset labels is the same, and the target label includes: if there are 3 users or groups contained in the preset record list, the corresponding labels are four, and the labels are respectively: when the similarity values are calculated, the similarity values are respectively 100%, 75%, 0 and 25%, according to practical analysis, the user or group corresponding to the tag "sports, football, westinghouse and queen horse" is similar to the interest of the user or group corresponding to the target tag, and the user or group corresponding to the tag "sports, football, westinghouse and queen horse" is the same as the interest of the user or group corresponding to the target tag, so that the user information to be matched is the user or group corresponding to the tag "sports, football, westinghouse and queen horse" and the user or group corresponding to the tag "sports, football, westinghouse and queen horse". For setting the preset threshold, the determination may be performed according to the number of actual tags, which is not limited in particular. Taking the above description as an example, the preset threshold at this time may be set to 75%.
Further, in step S30, when receiving the input selection instruction, the selected matching user is matched with the target user corresponding to the target chat text, and after the chat group is established, the method further includes: when the completion of the chat group creation is detected, issuing an authority instruction to the terminals contained in the chat group, and determining the state information of the chat group based on the received feedback information.
After the chat group creation is completed, the state information of the chat group may be set, for example, the chat group may be set to a state allowing others to enter directly or a state requiring agreement. All the user information which can be used for creating the group is recorded in the preset record list, wherein the user information comprises users or groups which can be directly matched and also comprises users or groups which need to be determined. When the chat group is created, the chat group creating device sends prompt information to all users in the created chat group to determine state information of the created chat group, wherein the specific prompt information can be: whether the aggregation can be performed by member agreement is required, and since the number of users in the group is a value which is not one, the aggregation can be performed by determining that the member agreement is required when the feedback information is that the agreed number reaches a certain percentage.
In the case of the present embodiment, the created chat group is visible, that is, the chat group which is allowed to be used as the user information to be matched after being queried by others, but whether to aggregate needs to be determined again. Thus, the safety of the chat group can be effectively ensured.
Further, referring to fig. 5, fig. 5 is a flowchart illustrating the steps for training a predictive model in one embodiment.
Specifically, the method further comprises:
step S80, when a model training instruction is received, receiving the input chat text to be classified and the classification label;
step S90, training chat texts to be classified and classification labels as input of a prediction model to be trained;
and step S100, when the fact that the prediction model to be trained starts to converge is detected, determining that the training of the prediction model to be trained is completed, and obtaining a trained prediction model.
And training the prediction model to be trained by acquiring a large number of chat texts to be trained and corresponding classification labels. Specifically, chat text to be classified and classification labels are used as model training inputs, then machine learning mode can automatically induce training, model training is determined to be completed when model convergence is detected, and then a trained prediction model can be obtained.
Further, the chat text to be trained and the corresponding classification labels are input into a prediction model to be trained for training, and the output of the prediction model to be trained is a classification rule corresponding to the chat text to be trained and the classification labels. Specifically, the process of training the prediction model to be trained through the chat text to be trained and the corresponding classification labels is the same as the prediction process of the prediction model to the target chat text, and firstly, through preprocessing the chat text to be trained, meaningless symbol information or other redundant information is mainly eliminated. For example, there may be some html tags on the corpus obtained using crawlers, these symbols should be redundant nonsensical information for text classification tasks, and can be culled out; the pre-processed chat text is then digitized, where the text is digitized in a variety of ways, such as: TF-IDF, BOW, one-Hot, distributed representation (word 2vec, glove), etc.; further, inputting the obtained numerical result and the classification label into a preset convolutional neural network, wherein the preset convolutional neural network comprises 13 convolutional layers, 13 excitation layers and 4 pooling layers, and performing characteristic extraction by performing operations such as convolution, excitation, pooling and the like on the numerical result and the classification label; finally, the parameter is adjusted through the full connection layer.
Based on the above process, the to-be-trained prediction model is continuously trained through the chat text to be trained and the corresponding classification labels, in the training process, when the convergence of the to-be-trained prediction model is detected, the current training can be judged to be completed, and the trained prediction model is obtained, specifically, judging that the to-be-trained prediction model converges can include but is not limited to the following cases: the training times reach the preset times, the training time reaches the preset time, the training loss function approaches zero, and the training loss function can be specifically set according to actual conditions.
The invention further provides a chat group creation device.
Referring to fig. 6, fig. 6 is a block diagram illustrating a structure of a chat group creation apparatus in an embodiment.
In one embodiment, the chat group creation device 60 includes a text acquisition module 61, a user determination module 62, and a group creation module 63.
A text obtaining module 61, configured to obtain a target chat text to be analyzed and processed when it is detected that the user matching function is triggered;
the user determining module 62 is configured to perform analysis and prediction on the target chat text based on a pre-trained prediction model, and obtain user information to be matched based on the obtained analysis and prediction result;
The group creation module 63 is configured to, when receiving an input selection instruction, match a matching user selected in the user information to be matched with a target user corresponding to the target chat text, and create a chat group.
In one embodiment, referring to fig. 7, fig. 7 is a block diagram of a chat group creation device according to another embodiment, and the chat group creation device 60 further includes:
a first determining module 64, configured to determine, when receiving instruction information for user matching sent by the first terminal, whether there is a second terminal that performs communication connection with the first terminal;
an instruction sending module 65, configured to send instruction information to a second terminal when it is determined that the second terminal that is in communication connection with the first terminal exists;
the function triggering module 66 is configured to trigger a user matching function when a determination instruction for allowing user matching sent by the second terminal is received within a preset time period; and the method is used for regarding the instruction information as an invalid instruction and sending corresponding prompt information to the first terminal when the second sent determination instruction which allows the user to be matched is not received within the preset time length.
In one embodiment, the instruction sending module 65 is specifically further configured to: and triggering a user matching function when the second terminal which is in communication connection with the first terminal is determined to be absent.
In one embodiment, referring to fig. 8, fig. 8 is a block diagram of a text obtaining module in one embodiment, where the text obtaining module 61 includes:
a time acquisition unit 611, configured to, when receiving a user matching instruction, acquire time information corresponding to the user matching instruction;
a text acquisition unit 612, configured to acquire, based on the time information, the chat text within the preset time period as the target chat text.
In one embodiment, referring to fig. 9, fig. 9 is a block diagram of a user determination module in one embodiment, the user determination module 62 includes:
a preprocessing unit 621, configured to preprocess the target chat text, so as to obtain a preprocessed target chat text;
a text input unit 622, configured to input the preprocessed target chat text into a preset convolutional neural network, so as to obtain a target label corresponding to the target chat text output by the convolutional neural network;
the user query unit 623 is configured to query and match in a preset record list based on the target tag, to obtain user information to be matched, where the preset record list records a correspondence between the user information and the tag information.
In one embodiment, the user query unit 623 includes:
A tag obtaining subunit 6231, configured to obtain a preset record list, and read first tag information corresponding to user information included in the preset record list;
a similarity calculation subunit 6232 configured to calculate a similarity value between the first tag information and the target tag;
a similarity comparison subunit 6233, configured to compare the similarity value with a preset threshold;
the user determining subunit 6234 is configured to determine, when the similarity value is greater than or equal to a preset threshold, that the user information associated with the first tag information is user information to be matched.
In one embodiment, referring to fig. 10, fig. 10 is a block diagram of a chat group creation device according to another embodiment, and the chat group creation device 60 further includes:
the information receiving module 67 is configured to receive the input chat text to be classified and the classification label when receiving the model training instruction;
the model training module 68 is configured to train the chat text to be classified and the classification label as input of the prediction model to be trained;
and the detection completion module 69 is configured to determine that the training of the prediction model to be trained is completed when it is detected that the prediction model to be trained begins to converge, so as to obtain a trained prediction model.
In one embodiment, a computer device is provided, the computer device comprising a memory and a processor, the memory having stored therein computer readable instructions that, when executed by the processor, cause the processor to perform the steps of:
when the user matching function is detected to be triggered, acquiring a target chat text to be analyzed and processed;
analyzing and predicting the target chat text based on a pre-trained prediction model, and acquiring user information to be matched based on an obtained analysis and prediction result;
when receiving an input selection instruction, matching the matched user selected in the user information to be matched with a target user corresponding to the target chat text, and establishing a chat group.
In one embodiment, the processor when executing the computer program further performs the steps of:
when receiving instruction information for user matching sent by a first terminal, determining whether a second terminal in communication connection with the first terminal exists or not;
when the second terminal in communication connection with the first terminal is determined to exist, sending instruction information to the second terminal;
triggering a user matching function when a determining instruction which is sent by the second terminal and allows user matching is received within a preset time period;
And when the second sent determining instruction allowing user matching is not received within the preset time, the instruction information is regarded as an invalid instruction, and corresponding prompt information is sent to the first terminal.
In one embodiment, the processor when executing the computer program further performs the steps of:
and triggering a user matching function when the second terminal which is in communication connection with the first terminal is determined to be absent.
In one embodiment, the processor when executing the computer program further performs the steps of:
when a user matching instruction is received, acquiring time information corresponding to the user matching instruction;
and acquiring the chat text in the preset time period based on the time information as a target chat text.
In one embodiment, the processor when executing the computer program further performs the steps of:
preprocessing the target chat text to obtain a preprocessed target chat text;
inputting the preprocessed target chat text into a preset convolutional neural network to obtain a target label corresponding to the target chat text output by the convolutional neural network;
and carrying out query matching in a preset record list based on the target label to obtain user information to be matched, wherein the preset record list records the corresponding relation between the user information and the label information.
In one embodiment, the processor when executing the computer program further performs the steps of:
acquiring a preset record list, and reading first tag information corresponding to user information contained in the preset record list;
calculating a similarity value between the first tag information and the target tag;
comparing the similarity value with a preset threshold value;
and when the similarity value is greater than or equal to a preset threshold value, determining the user information corresponding to the first tag information as the user information to be matched.
In one embodiment, the processor when executing the computer program further performs the steps of:
when a model training instruction is received, receiving the input chat text to be classified and the classification label;
training chat texts to be classified and classification labels as input of a prediction model to be trained;
when the convergence of the prediction model to be trained is detected, determining that the training of the prediction model to be trained is completed, and obtaining a trained prediction model.
In one embodiment, a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of:
When the user matching function is detected to be triggered, acquiring a target chat text to be analyzed and processed;
analyzing and predicting the target chat text based on a pre-trained prediction model, and acquiring user information to be matched based on an obtained analysis and prediction result;
when receiving an input selection instruction, matching the matched user selected in the user information to be matched with a target user corresponding to the target chat text, and establishing a chat group.
In one embodiment, the computer readable instructions, when executed by one or more processors, cause the one or more processors to further perform the steps of:
when receiving instruction information for user matching sent by a first terminal, determining whether a second terminal in communication connection with the first terminal exists or not;
when the second terminal in communication connection with the first terminal is determined to exist, sending instruction information to the second terminal;
triggering a user matching function when a determining instruction which is sent by the second terminal and allows user matching is received within a preset time period;
and when the second sent determining instruction allowing user matching is not received within the preset time, the instruction information is regarded as an invalid instruction, and corresponding prompt information is sent to the first terminal.
In one embodiment, the computer readable instructions, when executed by one or more processors, cause the one or more processors to further perform the steps of:
and triggering a user matching function when the second terminal which is in communication connection with the first terminal is determined to be absent.
In one embodiment, the computer readable instructions, when executed by one or more processors, cause the one or more processors to further perform the steps of:
when a user matching instruction is received, acquiring time information corresponding to the user matching instruction;
and acquiring the chat text in the preset time period based on the time information as a target chat text.
In one embodiment, the computer readable instructions, when executed by one or more processors, cause the one or more processors to further perform the steps of:
preprocessing the target chat text to obtain a preprocessed target chat text;
inputting the preprocessed target chat text into a preset convolutional neural network to obtain a target label corresponding to the target chat text output by the convolutional neural network;
and carrying out query matching in a preset record list based on the target label to obtain user information to be matched, wherein the preset record list records the corresponding relation between the user information and the label information.
In one embodiment, the computer readable instructions, when executed by one or more processors, cause the one or more processors to further perform the steps of:
acquiring a preset record list, and reading first tag information corresponding to user information contained in the preset record list;
calculating a similarity value between the first tag information and the target tag;
comparing the similarity value with a preset threshold value;
and when the similarity value is greater than or equal to a preset threshold value, determining the user information corresponding to the first tag information as the user information to be matched.
In one embodiment, the computer readable instructions, when executed by one or more processors, cause the one or more processors to further perform the steps of:
when a model training instruction is received, receiving the input chat text to be classified and the classification label;
training chat texts to be classified and classification labels as input of a prediction model to be trained;
when the convergence of the prediction model to be trained is detected, determining that the training of the prediction model to be trained is completed, and obtaining a trained prediction model.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM), comprising instructions for causing a terminal (which may be a mobile phone, a computer, a server or a network device, etc.) to perform the method according to the embodiments of the present invention.
While the embodiments of the present invention have been described above with reference to the drawings, the present invention is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and many modifications may be made thereto by those of ordinary skill in the art without departing from the spirit of the present invention and the scope of the appended claims, which are to be accorded the full scope of the present invention as defined by the following description and drawings, or by any equivalent structures or equivalent flow changes, or by direct or indirect application to other relevant technical fields.