CN113468297B

CN113468297B - Dialogue data processing method and device, electronic equipment and storage equipment

Info

Publication number: CN113468297B
Application number: CN202010236572.5A
Authority: CN
Inventors: 姜剑
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2020-03-30
Filing date: 2020-03-30
Publication date: 2024-02-27
Anticipated expiration: 2040-03-30
Also published as: CN113468297A

Abstract

The application discloses a dialogue data processing method, which comprises the following steps: acquiring dialogue data to be analyzed, and acquiring associated dialogue data corresponding to the dialogue data to be analyzed, wherein a preset correlation condition is met between the associated dialogue data and the dialogue data to be analyzed; according to the dialogue data to be analyzed and the associated dialogue data, obtaining target content characteristic information and target associated characteristic information corresponding to the dialogue data to be analyzed; and determining whether the dialogue data to be analyzed is used for consulting a problem according to the target content characteristic information and the target associated characteristic information. The method can conveniently and accurately obtain the dialogue data for consulting the problems.

Description

Dialogue data processing method and device, electronic equipment and storage equipment

Technical Field

The present invention relates to the field of computer technologies, and in particular, to a method and an apparatus for processing dialogue data, and an electronic device. The application also relates to a method and a device for obtaining the problem decision model and electronic equipment. The application also relates to a data pushing method and device and electronic equipment. The application also relates to a data display method and device and electronic equipment.

Background

With the continuous development of internet technology, online conversations are performed between customer service and clients, especially online conversations of multiple customer service and multiple clients are performed by a multi-user conversation group, so as to solve the problem that the clients may encounter when using enterprise products or enjoying enterprise services, and bring great convenience to enterprises and clients.

At present, when judging whether the dialogue data is the dialogue data for consulting the problems of product use or product improvement and the like, one common method is that in the process of dialogue between a client and enterprise customer service, the enterprise customer service manually determines whether the dialogue data is related to the problems, and if the dialogue data is related to the problems, a background technician answers or solves a production cost task list; of course, there is also a recognition model obtained by training corresponding keywords in advance, identifying related keywords in dialogue data of clients and enterprise clients, taking the dialogue data including the corresponding keywords as dialogue data which may have problems, providing the dialogue data to customer service personnel, and confirming the dialogue data by the customer service personnel, wherein the keywords used for training the recognition model are generally keywords or regular expressions including words such as "ask", "how", and the like.

According to the above description, in the current method for obtaining the dialogue data of the consultation problem from the original dialogue data of the consultation problem of the customer and the enterprise customer service, one common method is to manually participate, that is, the customer service needs to pay attention to the dialogue data at the moment, which is time-consuming and labor-consuming, and is inconvenient, especially when the dialogue is generated in a multi-user dialogue group, because there may be a plurality of dialogue problems of different customers at the same time, the difficulty that will be caused by the customer service to judge whether the dialogue data is used for the consultation service is greatly increased; in addition, one method, which has been widely used at present, is to identify whether dialogue data is used for a consultation problem based on keywords, but if a client does not use related keywords at the time of consultation of the problem, dialogue data for the consultation problem cannot be accurately obtained. From this, it is known that the prior art has a problem of being inconvenient and accurate when acquiring dialogue data for consultation problems from original dialogue data.

Disclosure of Invention

The embodiment of the application provides a dialogue data processing method, which aims to solve the problem that the prior art is inconvenient and accurate when dialogue data for consultation problems are acquired from original dialogue data.

The embodiment of the application provides a dialogue data processing method, which comprises the following steps:

acquiring dialogue data to be analyzed, and acquiring associated dialogue data corresponding to the dialogue data to be analyzed, wherein a preset correlation condition is met between the associated dialogue data and the dialogue data to be analyzed; according to the dialogue data to be analyzed and the associated dialogue data, obtaining target content characteristic information and target associated characteristic information corresponding to the dialogue data to be analyzed; and determining whether the dialogue data to be analyzed is used for consulting a problem according to the target content characteristic information and the target associated characteristic information.

Optionally, the method further comprises: acquiring original dialogue data; the step of acquiring dialogue data to be analyzed and acquiring associated dialogue data corresponding to the dialogue data to be analyzed comprises the following steps: and acquiring the dialogue data to be analyzed from the original dialogue data, and acquiring the associated dialogue data corresponding to the dialogue data to be analyzed.

Optionally, the acquiring the dialogue data to be analyzed from the original dialogue data and the acquiring the associated dialogue data corresponding to the dialogue data to be analyzed include:

And acquiring the dialogue data to be analyzed from the original dialogue data, and acquiring first associated dialogue data corresponding to the dialogue data to be analyzed, wherein the position of the dialogue data to be analyzed in the original dialogue data is adjacent to the position of the first associated dialogue data, and the time information corresponding to the dialogue data to be analyzed is later than the time information corresponding to the first associated dialogue data.

Optionally, the target content characteristic information corresponding to the dialogue data to be analyzed is obtained by the following method:

inputting the dialogue data to be analyzed into a first recognition model, and acquiring first content characteristic information corresponding to the dialogue data to be analyzed, wherein the first recognition model is used for recognizing whether the dialogue data to be analyzed is used for consulting a first problem;

inputting the first associated dialogue data into a second recognition model, and acquiring second content characteristic information corresponding to the second dialogue data, wherein the second recognition model is used for recognizing whether the first associated dialogue data is dialogue data for ending a dialogue corresponding to consultation of a second problem;

And obtaining the target content characteristic information according to the first content characteristic information and the second content characteristic information.

Optionally, the target associated feature information corresponding to the dialogue data to be analyzed is obtained by the following method:

acquiring a first user type corresponding to a speaker of the dialogue data to be analyzed;

judging whether the speaker of the dialogue data to be analyzed is the same as the speaker of the first associated dialogue data or not, and obtaining a first judgment result;

acquiring a time interval between the dialogue data to be analyzed and the dialogue data to be analyzed of first associated dialogue data as first time interval data;

and obtaining the target associated feature information according to the first user type, the first judging result and the first time interval data.

Optionally, the determining whether the dialogue data to be analyzed is used for the consultation problem according to the target content feature information and the target association feature information includes:

inputting the target content characteristic information and the target associated characteristic information into a problem decision model to obtain a target decision result, wherein the problem decision model is used for determining whether the dialogue data to be analyzed is used for consulting a problem or not according to the obtained target content characteristic information and the target associated characteristic information, and the target decision result is used for identifying whether the dialogue data to be analyzed is used for consulting the problem or not.

Optionally, the acquiring the dialogue data to be analyzed from the original dialogue data, and acquiring the associated dialogue data corresponding to the dialogue data to be analyzed, further includes:

and acquiring second associated dialogue data corresponding to the dialogue data to be analyzed from the original dialogue data, wherein the dialogue data to be analyzed and the second associated dialogue data are separated by a preset amount of dialogue data, and the time information corresponding to the dialogue data to be analyzed is later than the time information corresponding to the associated dialogue data.

Optionally, the target content characteristic information corresponding to the dialogue data to be analyzed is further obtained by the following method:

acquiring third content characteristic information corresponding to the second associated dialogue data;

and obtaining the target content characteristic information according to the first content characteristic information, the second content characteristic information and the third content characteristic information.

Optionally, the target associated feature information corresponding to the dialogue data to be analyzed is further obtained by the following method:

acquiring a second user type corresponding to a speaker of the second associated dialogue data;

Judging whether the speaker of the dialogue data to be analyzed is the same as the speaker of the second associated dialogue data or not, and obtaining a second judgment result;

acquiring a time interval between the dialogue data to be analyzed and the second associated dialogue data as second time interval data;

and obtaining the target associated characteristic information according to the first user type, the first judging result, the first time interval data, the second user type, the second judging result and the second time interval data.

Optionally, the method further comprises: and if the dialogue data to be analyzed is determined to be used for consulting the problem, generating task data which corresponds to the dialogue data to be analyzed and is used for processing the problem.

Optionally, the method is applied to a server, and the generating task data corresponding to the dialogue data to be analyzed and used for processing the problem includes:

providing the dialogue data to be analyzed to a client;

receiving a data request message sent by a client for generating the task data;

and generating the task data according to the data request message.

Optionally, the method is applied to a client, and the generating task data corresponding to the dialogue data to be analyzed and used for processing the problem includes:

receiving the dialogue data to be analyzed provided by a server;

generating a data request message for generating the task data according to the dialogue data to be analyzed;

and sending the data request message to the server.

Optionally, the problem decision model is obtained by the following method:

acquiring relevant dialogue data for consultation problems from historical dialogue data as positive sample dialogue data, and acquiring relevant dialogue data which is not used for consultation problems as negative sample dialogue data, wherein the preset correlation conditions are met between the dialogue data in the positive sample dialogue data and the dialogue data in the negative sample dialogue data;

training to obtain the problem decision model by using the positive sample dialogue data and the negative sample data as sample data;

the sample data comprises sample dialogue data to be analyzed and associated dialogue data meeting the preset correlation condition between the sample dialogue data to be analyzed and the problem decision model is used for determining whether the sample dialogue data to be analyzed is used for consulting a problem according to input sample content characteristic information and sample associated characteristic information corresponding to the sample dialogue data to be analyzed.

Optionally, the obtaining the target content feature information according to the first content feature information and the second content feature information includes:

acquiring first symbol identification information and second symbol identification information according to the dialogue data to be analyzed, wherein the first symbol identification information is used for identifying whether an exclamation mark exists in the dialogue data to be analyzed, and the second symbol identification information is used for identifying whether a question mark exists in the dialogue data to be analyzed;

acquiring third symbol identification information and fourth symbol identification information according to the first associated dialogue data, wherein the third symbol identification information is used for identifying whether an exclamation mark exists in the first associated dialogue data, and the fourth symbol identification information is used for identifying whether a question mark exists in the first associated dialogue data;

and obtaining the target content characteristic information according to the first content characteristic information, the second content characteristic information, the first symbol identification information, the second symbol identification information, the third symbol identification information and the fourth symbol identification information.

Optionally, the first content feature information includes: the first result identification information is used for identifying whether the dialogue data to be analyzed corresponds to the first problem or not, and the first score information is used for indicating the accuracy degree of the first result identification information;

The second content feature information includes: and second result identification information for identifying whether the first associated session data is session data for ending a session corresponding to the second question, and second score information for indicating a degree of correctness of the second result identification information.

Optionally, the acquiring third content feature information corresponding to the second association dialogue data includes:

obtaining fifth symbol identification information and sixth symbol identification information according to the second associated dialogue data, wherein the fifth symbol identification information is used for identifying whether an exclamation mark exists in the second associated dialogue data, and the sixth symbol identification information is used for identifying whether a question mark exists in the second associated dialogue data;

and obtaining the third content characteristic information according to the fifth symbol identification information and the sixth symbol identification information.

Optionally, the original dialogue data includes text data, audio data, and video data.

The embodiment of the application also provides another dialogue data processing method, which comprises the following steps:

acquiring original dialogue data from a multi-user dialogue group for providing a target product consultation service for a user;

Acquiring current dialogue data from the original dialogue data as dialogue data to be analyzed, and acquiring associated dialogue data corresponding to the dialogue data to be analyzed, wherein a preset correlation condition is met between the associated dialogue data and the dialogue data to be analyzed;

according to the dialogue data to be analyzed and the associated dialogue data, obtaining target content characteristic information and target associated characteristic information corresponding to the dialogue data to be analyzed;

and determining whether the dialogue data to be analyzed is dialogue data for consulting the use problem of the target product according to the target content characteristic information and the target associated characteristic information.

The embodiment of the application also provides a method for obtaining the problem decision model, which comprises the following steps:

training to obtain a problem decision model by using the positive sample dialogue data and the negative sample data as sample data;

The embodiment of the application also provides a data pushing method, which is applied to the server and comprises the following steps:

acquiring original dialogue data sent by a client; acquiring target problem list data corresponding to the original dialogue data, wherein the target problem list data is data which is obtained by using the dialogue data processing method and comprises at least one piece of dialogue data for consulting a problem; and pushing the target problem list data to the client.

Optionally, the method further comprises: acquiring a data request message sent by the client for acquiring target task data, wherein the target task data is data which corresponds to dialogue data for consulting a problem in the target problem list data and is used for processing the problem; generating the target task data according to the target problem list data; and pushing the target task data to the client.

The embodiment of the application also provides a data display method, which is applied to the client and comprises the following steps: acquiring original dialogue data; sending the original dialogue data to a server; and receiving target problem list data pushed by the server and corresponding to the original dialogue data, wherein the target problem list data is data which is obtained by the server through the dialogue data processing method and contains at least one piece of dialogue data for consulting a problem.

Optionally, the method further comprises: sending a data request message for obtaining target task data to the server, wherein the target task data is data which corresponds to dialogue data for consulting a problem in the target problem list data and is used for processing the problem; receiving the target task data pushed by the server; and displaying the target task data.

The embodiment of the application also provides a dialogue data processing device, which comprises:

a data acquisition unit, configured to acquire dialogue data to be analyzed from the original dialogue data, and acquire associated dialogue data corresponding to the dialogue data to be analyzed, where a preset correlation condition is satisfied between the associated dialogue data and the dialogue data to be analyzed;

The feature information acquisition unit is used for acquiring target content feature information and target associated feature information corresponding to the dialogue data to be analyzed according to the dialogue data to be analyzed and the associated dialogue data;

and the determining unit is used for determining whether the dialogue data to be analyzed is used for consulting the problem according to the target content characteristic information and the target associated characteristic information.

The embodiment of the application also provides electronic equipment, which comprises:

a processor;

a memory for storing a program of a dialogue data processing method, the device being powered on and executing the program of the dialogue data processing method by the processor, performing the steps of:

acquiring dialogue data to be analyzed, and acquiring associated dialogue data corresponding to the dialogue data to be analyzed, wherein a preset correlation condition is met between the associated dialogue data and the dialogue data to be analyzed;

and determining whether the dialogue data to be analyzed is used for consulting a problem according to the target content characteristic information and the target associated characteristic information.

The embodiment of the application also provides another dialogue data processing device, which comprises:

an original dialogue data acquisition unit that acquires original dialogue data from a multi-user dialogue group for providing a target product consultation service for a user;

a data acquisition unit, configured to acquire current dialogue data as dialogue data to be analyzed from the original dialogue data, and acquire associated dialogue data corresponding to the dialogue data to be analyzed, where a preset correlation condition is satisfied between the associated dialogue data and the dialogue data to be analyzed;

and the determining unit is used for determining whether the dialogue data to be analyzed is dialogue data for consulting the use problem of the target product according to the target content characteristic information and the target associated characteristic information.

The embodiment of the application also provides another electronic device, which comprises:

a processor;

The embodiment of the application also provides a device for obtaining the problem decision model, which comprises the following steps:

a sample data obtaining unit, configured to obtain relevant dialogue data for consultation problems from historical dialogue data as positive sample dialogue data, and obtain relevant dialogue data not for consultation problems as negative sample dialogue data, where the preset correlation condition is satisfied between dialogue data in the positive sample dialogue data and dialogue data in the negative sample dialogue data;

The training unit is used for training to obtain a problem decision model by using the positive sample dialogue data and the negative sample data as sample data;

a processor;

a memory for storing a program of a method of obtaining a problem decision model, the apparatus being powered on and executing the program of the method of obtaining a problem decision model by the processor, and performing the steps of:

Compared with the prior art, the application has the following advantages:

the embodiment of the application provides a dialogue data processing method, which comprises the following steps: acquiring dialogue data to be analyzed, and acquiring associated dialogue data corresponding to the dialogue data to be analyzed, wherein a preset correlation condition is met between the associated dialogue data and the dialogue data to be analyzed; according to the dialogue data to be analyzed and the associated dialogue data, obtaining target content characteristic information and target associated characteristic information corresponding to the dialogue data to be analyzed; and determining whether the dialogue data to be analyzed is used for consulting a problem according to the target content characteristic information and the target associated characteristic information. According to the method, whether the dialogue data are used for consulting the problem is not needed in a manual mode or according to the keyword information, and the dialogue data to be analyzed and the associated dialogue data meeting the preset correlation condition between the dialogue data to be analyzed are obtained, the multidimensional characteristic information corresponding to the dialogue data to be analyzed in the dialogue process is obtained, and whether the dialogue data to be analyzed are used for consulting the problem is determined according to the multidimensional characteristic information, so that the dialogue data related to the problem can be conveniently and accurately obtained.

Drawings

Fig. 1-a is a schematic diagram of a first application scenario of a method for processing dialogue data according to a first embodiment of the present application.

Fig. 1-B is a schematic diagram of a second application scenario of a session data processing method according to the first embodiment of the present application.

Fig. 1-C is a schematic diagram of a third application scenario of a session data processing method according to the first embodiment of the present application.

Fig. 2 is a flowchart of a dialogue data processing method according to the first embodiment of the present application.

Fig. 3 is a flowchart of another dialogue data processing method according to the second embodiment of the present application.

Fig. 4 is a flowchart of a method for obtaining a problem decision model according to a third embodiment of the present application.

Fig. 5 is a schematic diagram of a dialogue data processing device according to a fourth embodiment of the present application.

Fig. 6 is a schematic diagram of an electronic device according to a fifth embodiment of the present application.

Fig. 7 is a schematic diagram of another dialogue data processing device according to the sixth embodiment of the present application.

Fig. 8 is a schematic diagram of an apparatus for obtaining a problem decision model according to an eighth embodiment of the present application.

Fig. 9 is a flowchart of a data pushing method according to a tenth embodiment of the present application.

Fig. 10 is a flowchart of a data display method according to an eleventh embodiment of the present application.

Fig. 11 is a schematic diagram of a data pushing device according to a twelfth embodiment of the present application.

Fig. 12 is a schematic diagram of a data display device according to a fourteenth embodiment of the present application.

Detailed Description

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. This application is, however, susceptible of embodiment in many other ways than those herein described and similar generalizations can be made by those skilled in the art without departing from the spirit of the application and the application is therefore not limited to the specific embodiments disclosed below.

In order to enable those skilled in the art to better understand the scheme of the present application, a detailed description is given below of a specific application scenario of an embodiment thereof based on the session data processing method provided in the present application. The method can be applied to a scenario of interaction between a client and a server, as shown in fig. 1-a, which is a schematic diagram of a first application scenario of a session data processing method provided in a first embodiment of the present application.

In a specific implementation, the method may be applied to analyze dialogue data in a multi-user dialogue group providing a target product consultation service for a user, and provide dialogue data for consultation problems, for example, for consultation of a use problem of the target product, for the enterprise customer service to respond to a client problem in time, specifically, after the client obtains the original dialogue data, send the original dialogue data to the server; the method comprises the steps that a server side obtains dialogue data to be analyzed from received original dialogue data, and obtains associated dialogue data corresponding to the dialogue data to be analyzed, wherein the associated dialogue data and the dialogue data to be analyzed meet preset correlation conditions; then, the server obtains target content characteristic information and target associated characteristic information corresponding to the dialogue data to be analyzed according to the dialogue data to be analyzed and the associated dialogue data; determining whether the dialogue data to be analyzed is used for consulting the problem according to the target content characteristic information and the target associated characteristic information, and if the dialogue data to be analyzed is used for consulting the problem, providing the dialogue data to be analyzed to a client by a server; after the client obtains the dialogue data to be analyzed, the dialogue data to be analyzed is displayed, or a notification message aiming at the dialogue data to be analyzed can be generated, or task data such as a task list aiming at the dialogue data to be analyzed can be generated, so that enterprise customer service and related technical staff in the background can respond to the problems in the dialogue data to be analyzed in time, the clients can respond in time, and the client experience is improved.

As shown in fig. 1-B, which is a schematic diagram of a second application scenario of a dialogue data processing method according to the first embodiment of the present application, as can be seen from fig. 1-B, when determining to-be-analyzed dialogue data "speaker C (15:30:00): you get a consultation that i have a question … …? "is dialogue data for consultation questions," the client can display the dialogue data for the customer service of the enterprise to view. In addition, as shown in fig. 1-C, which is a schematic diagram of a third application scenario of a dialogue data processing method provided in the first embodiment of the present application, according to the content of fig. 1-C, it can be known that an enterprise customer service can generate a task sheet for a corresponding problem in the dialogue data to be analyzed according to the dialogue data to be analyzed displayed by a client.

In addition, if the server corresponding to the dialogue platform for the dialogue between the client and the enterprise customer service is self-developed by the enterprise, that is, the server does not need the client to send the original dialogue data, but can automatically obtain the original dialogue data in the process of the dialogue between the client and the enterprise customer service, the method can also be independently applied to the server. Specifically, after the server obtains the original dialogue data of the dialogue between the client and the enterprise customer service, the server obtains the current dialogue data from the original dialogue data in real time as dialogue data to be analyzed and obtains the associated dialogue data corresponding to the dialogue data to be analyzed, and obtains the target content characteristic information and the target associated characteristic information corresponding to the dialogue data to be analyzed, then determines whether the dialogue data to be analyzed is used for consultation problems according to the target content characteristic information and the target associated characteristic information, if so, the server provides the dialogue data to be analyzed for the enterprise customer service, or the server directly generates corresponding task data, such as a task list, according to a background technician corresponding to the problem corresponding to the dialogue data to be analyzed, and pushes the task data to the background technician.

Of course, the method can also be independently applied to a client used by enterprise customer service, after the client obtains the original dialogue data of the dialogue between the client and the enterprise customer service, the original dialogue data is not required to be sent to a server, the dialogue data to be analyzed and the associated dialogue data are directly obtained from the original dialogue data, whether the dialogue data to be analyzed are used for consultation problems or not is determined through the obtained target content characteristic information and the target associated characteristic information corresponding to the dialogue data to be analyzed, and if yes, the client directly generates a notification message aiming at the dialogue data to be analyzed so as to be checked by the enterprise customer service and respond timely.

The client may be a mobile terminal device, such as a mobile phone, a tablet computer, or a common computer device. The server is generally referred to as a server, and the server may be a physical server or a cloud server, which is not limited herein.

The target product may be an application program developed by an enterprise and used for providing an application service for a user, and of course, the target product may also be a physical product or other products providing virtual services, for example, the target product may be an application program in a mobile terminal, may be a physical product such as an article of daily use or a vehicle with a physical feature, or may be a virtual service providing a communication fee recharging service, etc.

It should be noted that the above application scenario is merely a specific embodiment of the method for processing dialogue data provided in the first embodiment of the present application, and the purpose of the above application scenario embodiment is to facilitate understanding the method provided in the first embodiment of the present application, and is not limited to the method.

Fig. 2 is a flowchart of a method for processing dialogue data according to the first embodiment of the present application, and the method is described in detail below with reference to fig. 2.

Step S201, acquiring dialogue data to be analyzed from the original dialogue data, and acquiring associated dialogue data corresponding to the dialogue data to be analyzed, wherein a preset correlation condition is satisfied between the associated dialogue data and the dialogue data to be analyzed.

The dialogue data to be analyzed may be any piece of dialogue data in the original dialogue data, and of course, if the server side can acquire the original dialogue data in real time, the dialogue data to be analyzed generally refers to dialogue data generated in real time, namely current dialogue data.

The original dialogue data is data comprising at least one piece of dialogue data sent by at least one user, the content of the original dialogue data can be at least one of text data, image data, audio data and video data, and the original dialogue data can be text or image data sent by the user through an instant messaging application program or voice or video dialogue data sent by the mobile terminal device.

It should be noted that, in the original dialogue data, one piece of dialogue data refers to content that is sent out by a single user at a time in a unit time, for example, when the dialogue data is text data, one piece of dialogue data may be a text that is sent out by a single user at a time in a unit time, or may be at least one piece of text that is sent out by a single user at a time in a unit time.

That is, before acquiring the dialogue data to be analyzed and acquiring the associated dialogue data corresponding to the dialogue data to be analyzed, the method provided in the first embodiment of the present application further includes: acquiring original dialogue data; the step of acquiring dialogue data to be analyzed and acquiring associated dialogue data corresponding to the dialogue data to be analyzed comprises the following steps: and acquiring the dialogue data to be analyzed from the original dialogue data, and acquiring the associated dialogue data corresponding to the dialogue data to be analyzed.

In practice, in single chat session data, that is, session data in which clients and enterprise customer services individually perform a session, it is relatively simple to determine whether a certain session data is used for a consultation problem; in the first embodiment of the present application, if no special explanation exists, the original dialogue data is used as the dialogue data from the multi-user dialogue group providing the target product consultation service for the user to explain the dialogue data processing method.

For example, the original dialog data may be "1, speaker a (10:00:00): the problem has been solved; 2. speaker B (10:10:00): if other problems exist, the method can feed back to me; 3. speaker C (15:30:00): you get a consultation that i have a question … …? … … ", wherein speaker A and speaker C are clients and speaker B is enterprise customer service.

The obtaining of the original dialogue data may be that after the client confirms via the client or the enterprise customer service, the original dialogue data stored in the client is sent to the server when the client and the enterprise customer service are in dialogue; of course, after confirmation by the customer or the enterprise customer service, the server side may obtain the original dialogue data by itself, and use the dialogue processing method to confirm whether the original dialogue data has the dialogue data for consulting the problem.

The associated dialogue data comprises first associated dialogue data, wherein the position of the first associated dialogue data in the original dialogue data is adjacent to the position of dialogue data to be analyzed, and time information corresponding to the first associated dialogue data is earlier than time information corresponding to the dialogue data to be analyzed. That is, the first associated dialogue data may be the last dialogue data of the dialogue data to be analyzed, that is, the acquiring dialogue data to be analyzed from the original dialogue data, and the acquiring associated dialogue data corresponding to the dialogue data to be analyzed, including: and acquiring the dialogue data to be analyzed from the original dialogue data, and acquiring first associated dialogue data corresponding to the dialogue data to be analyzed, wherein the position of the dialogue data to be analyzed in the original dialogue data is adjacent to the position of the first associated dialogue data, and the time information corresponding to the dialogue data to be analyzed is later than the time information corresponding to the first associated dialogue data.

It should be noted that, in the implementation, the dialogue data to be analyzed may be at least one piece of dialogue data, and the associated dialogue data may also be at least one piece of dialogue data; in the first embodiment of the present application, in order to reduce the computational complexity and increase the speed of obtaining the final decision result, the dialogue data to be analyzed and the associated dialogue data are only one piece of dialogue data in the original dialogue data.

For example, for original dialog data "1, speaker a (10:00:00): the problem has been solved; 2. speaker B (10:10:00): if other problems exist, the method can feed back to me; 3. speaker C (15:30:00): you get a consultation that i have a question … …? If the dialogue data to be analyzed is the 3 rd dialogue data, the first associated dialogue data is the 2 nd dialogue data; and if the dialogue data to be analyzed is the 2 nd dialogue data, the first associated dialogue data is the 1 st dialogue data.

In addition, in order to further increase the feature information corresponding to the dialogue data to be analyzed in more dimensions, the associated dialogue data further includes second associated dialogue data, wherein the second associated dialogue data is dialogue data spaced by a predetermined amount between the dialogue data to be analyzed, and time information corresponding to the second associated dialogue data is earlier than time information corresponding to the dialogue data to be analyzed, that is, the second associated dialogue data may be the last two sentences of dialogue data of the dialogue data to be analyzed, that is, the second associated dialogue data is spaced by one piece of the first associated dialogue data from the dialogue data to be analyzed. Thus, the obtaining the dialogue data to be analyzed from the original dialogue data, and obtaining the associated dialogue data corresponding to the dialogue data to be analyzed, further includes: and acquiring second associated dialogue data corresponding to the dialogue data to be analyzed from the original dialogue data, wherein the dialogue data to be analyzed and the second associated dialogue data are separated by a preset amount of dialogue data, and the time information corresponding to the dialogue data to be analyzed is later than the time information corresponding to the associated dialogue data.

For example, for original dialog data "1, speaker a (10:00:00): the problem has been solved; 2. speaker B (10:10:00): if other problems exist, the method can feed back to me; 3. speaker C (15:30:00): you get a consultation that i have a question … …? If the dialogue data to be analyzed is the 3 rd dialogue data, the second associated dialogue data is the 1 st dialogue data. It should be noted that, the predetermined number is 1, and the predetermined number may be set to other values as required in the implementation.

In addition, there may be a plurality of dialogue data in the original dialogue data, so the original dialogue data may be split into at least one pair of dialogue data pairs composed of opposite dialogue data to be analyzed and associated dialogue data according to the above method, and then multidimensional feature information, such as content feature information and associated feature information, corresponding to the dialogue data to be analyzed in each dialogue data pair may be acquired according to each dialogue data pair.

After step S201, step S202 is performed to obtain target content feature information and target associated feature information corresponding to the dialogue data to be analyzed according to the dialogue data to be analyzed and the associated dialogue data.

After at least one pair of dialogue data pairs consisting of opposite dialogue data to be analyzed and associated dialogue data are obtained through the steps, target content characteristic information and target associated characteristic information corresponding to the dialogue data to be analyzed can be constructed according to the dialogue data pairs, namely according to the dialogue data to be analyzed and the associated dialogue data, and serve as multidimensional characteristic information corresponding to the dialogue data to be analyzed.

In order to more clearly illustrate the above steps, a description will be given of a first recognition model and a second recognition model which are required to be used in the subsequent processing.

The first recognition model and the second recognition model are convolutional neural network models obtained through pre-training.

The first recognition model is used for recognizing whether the input content is used for consulting the first problem, and the output result of the first recognition model comprises first result identification information for identifying whether the input content is dialogue data used for consulting the first problem and first score information which corresponds to the first result identification information and is used for representing the accuracy degree of the first result identification information.

For example, the first result identification information may be 0 or 1, wherein 0 indicates that its input content is not dialogue data for referring to the first question, and 1 indicates that its input content is dialogue data for referring to the first question; the range of the first score information may be 0 to 1, and if the value of the first score information is close to 1, the accuracy of the first result identification information is higher.

And a second recognition model for recognizing whether the input content thereof is dialogue data for ending a dialogue for consulting a second question correspondence, the output result of the second recognition model including second result identification information for identifying whether the input content thereof is dialogue data for ending a dialogue for consulting the second question correspondence, and second score information corresponding to the second result identification information for indicating a degree of correctness of the second result identification information.

For example, the second result identification information may be 0 or 1, wherein 0 indicates that its input content is not dialogue data for ending a dialogue corresponding to the second question, and 1 indicates that its input content is dialogue data for ending a dialogue corresponding to the second question; the range page of the second score information may be 0 to 1, and if the value of the second score information approaches 1, it indicates that the accuracy of the second result identification information is high.

It should be noted that the first problem and the second problem are different problems, so the first recognition model and the second recognition model are used, because generally, in the original dialogue data, a piece of previous dialogue data before the dialogue data for consulting a new problem is the dialogue data for ending the dialogue of the previous problem, and in the first embodiment of the present application, when determining whether the dialogue data to be analyzed is the dialogue data for consulting a problem, the multidimensional feature information corresponding to the dialogue data to be analyzed is constructed by the dialogue data pair including the dialogue data to be analyzed and the first associated dialogue data, so if the first associated dialogue data is the dialogue data for ending the dialogue of the previous problem, the probability that the dialogue data to be analyzed is the dialogue data for consulting a new problem will be higher, and therefore, by identifying the dialogue data to be analyzed and the first associated dialogue data by the first recognition model and the second recognition model respectively, more accurate feature information for increasing the final decision result can be obtained.

In the first embodiment of the present application, the first recognition model and the second recognition model are text convolutional neural network models (TextCNN, text Convolutional Neural Networks), so if the dialogue data to be analyzed and the first associated dialogue data are image data, audio data or video data, they may be converted into corresponding text contents in advance, or other classification models for different formats may be trained in advance as the first recognition model and the second recognition model according to the difference in formats of the dialogue data to be analyzed and the first associated dialogue data.

The method for obtaining the first recognition model comprises the following steps: acquiring relevant dialogue data for consultation problems from historical dialogue data as positive sample dialogue data, and acquiring relevant dialogue data which is not used for consultation problems as negative sample dialogue data, wherein the preset correlation conditions are met between the dialogue data in the positive sample dialogue data and the dialogue data in the negative sample dialogue data; training to obtain the first recognition model by using the sample dialogue data to be analyzed included in the positive sample dialogue data and the negative sample data as first sample data.

The method for obtaining the second recognition model comprises the following steps: acquiring relevant dialogue data for consultation problems from historical dialogue data as positive sample dialogue data, and acquiring relevant dialogue data which is not used for consultation problems as negative sample dialogue data, wherein the preset correlation conditions are met between the dialogue data in the positive sample dialogue data and the dialogue data in the negative sample dialogue data; and training to obtain the second recognition model by using the first correlation sample dialogue data corresponding to the sample dialogue data to be analyzed, which is included in the positive sample dialogue data and the negative sample data, as second sample data.

The specific training process of the network model is not described in detail herein because of the detailed description in the prior art.

The first recognition model and the second recognition model required to be used in the subsequent processing are described in detail above, and how to obtain the target content feature information and the target associated feature information corresponding to the dialogue data to be analyzed according to the dialogue data to be analyzed and the associated dialogue data by using the first recognition model and the second recognition model obtained by the training in advance is described in detail below.

It should be noted that, when only the first associated dialogue data corresponding to the dialogue data to be analyzed is obtained in the above-described step S201, the target content feature information corresponding to the dialogue data to be analyzed may be obtained by: inputting the dialogue data to be analyzed into a first recognition model, and acquiring first content characteristic information corresponding to the dialogue data to be analyzed, wherein the first recognition model is used for recognizing whether the dialogue data to be analyzed is used for consulting a first problem; inputting the first associated dialogue data into a second recognition model, and acquiring second content characteristic information corresponding to the second dialogue data, wherein the second recognition model is used for recognizing whether the first associated dialogue data is dialogue data for ending a dialogue corresponding to consultation of a second problem; and obtaining the target content characteristic information according to the first content characteristic information and the second content characteristic information.

Wherein the first content feature information includes: the first result identification information is used for identifying whether the dialogue data to be analyzed corresponds to the first problem or not, and the first score information is used for indicating the accuracy degree of the first result identification information; the second content feature information includes: and second result identification information for identifying whether the first associated session data is session data for ending a session corresponding to the second question, and second score information for indicating a degree of correctness of the second result identification information.

For example, the dialogue data to be analyzed is "speaker C (15:30:00): you get a consultation that i have a question … …? "the first associated dialog data is" speaker B (10:10:00): if other problems can be fed back to me, inputting dialogue data to be analyzed into a first recognition model, and acquiring first result identification information of 1 and first score information of 0.95, namely first content characteristic information (1,0.95); inputting the first associated dialogue data into the second recognition model, and acquiring the second result identification information as 1, wherein the second score information as 0.98, namely the second content characteristic information as 1,0.98; then, from the first content feature information and the second content feature information, target content feature information corresponding to the dialogue data to be analyzed can be obtained as (1,0.95,1,0.98).

In addition, in order to increase the accuracy of the final decision result, feature information of other dimensions corresponding to the dialogue data to be analyzed can be further acquired, for example, symbol information in the dialogue data to be analyzed and the first associated dialogue data can be acquired, for example, when the dialogue data is in the form of exclamation marks, namely "+|! At the end, the dialogue data is usually used for showing thank you, ending the dialogue of the last question; when dialogue data is given a question mark, i.e., "? At the end, it is generally indicated that the dialogue data is dialogue data for consulting a new question, and therefore, when the target content feature information is obtained based on the first content feature information and the second content feature information, the dialogue data to be analyzed and the sign information in the first associated dialogue data can be further obtained as contents in the target content feature information.

Specifically, the obtaining the target content feature information according to the first content feature information and the second content feature information includes: acquiring first symbol identification information and second symbol identification information according to the dialogue data to be analyzed, wherein the first symbol identification information is used for identifying whether an exclamation mark exists in the dialogue data to be analyzed, and the second symbol identification information is used for identifying whether a question mark exists in the dialogue data to be analyzed; acquiring third symbol identification information and fourth symbol identification information according to the first associated dialogue data, wherein the third symbol identification information is used for identifying whether an exclamation mark exists in the first associated dialogue data, and the fourth symbol identification information is used for identifying whether a question mark exists in the first associated dialogue data; and obtaining the target content characteristic information according to the first content characteristic information, the second content characteristic information, the first symbol identification information, the second symbol identification information, the third symbol identification information and the fourth symbol identification information.

For example, the dialogue data to be analyzed is "speaker C (15:30:00): you get a consultation that i have a question … …? "the first associated dialog data is" speaker B (10:10:00): preferably, if there are other questions that can be fed back to me ", the first content feature information is obtained in the above process (1,0.95), and the second content feature information is obtained in the above process (1,0.98); in the dialogue data to be analyzed, a question mark exists, and an exclamation mark does not exist, so that the first symbol identification information is 0 and the second symbol identification information is 1 can be obtained from the dialogue data to be analyzed; from the first associated session data, the third symbol identification information is 0, and the fourth symbol identification information is 0, and the target content feature information is 1,0.95,1,0.98,0,1,0,0. It should be noted that, here, the symbol 0 is not present, the symbol 1 is present, and other values or forms may be used for the implementation, which is not described herein.

It should be noted that, if in the above step S201, the second associated dialogue data corresponding to the dialogue data to be analyzed is also obtained, that is, in order to further increase the accuracy of the decision result, the content feature information of more dimensions corresponding to the dialogue data to be analyzed may also be obtained through the second associated dialogue data, that is, the target content feature information corresponding to the dialogue data to be analyzed may also be obtained by: acquiring third content characteristic information corresponding to the second associated dialogue data; and obtaining the target content characteristic information according to the first content characteristic information, the second content characteristic information and the third content characteristic information.

The obtaining third content characteristic information corresponding to the second associated dialogue data includes: obtaining fifth symbol identification information and sixth symbol identification information according to the second associated dialogue data, wherein the fifth symbol identification information is used for identifying whether an exclamation mark exists in the second associated dialogue data, and the sixth symbol identification information is used for identifying whether a question mark exists in the second associated dialogue data; and obtaining the third content characteristic information according to the fifth symbol identification information and the sixth symbol identification information.

For example, the dialogue data to be analyzed is "speaker C (15:30:00): you get a consultation that i have a question … …? "the first associated dialog data is" speaker B (10:10:00): preferably, if there are other questions that can be fed back to me again, the second associated dialogue data is "speaker a (10:00:00): the problem is solved ", the first content characteristic information is (1,0.95) obtained in the above process, and the second content characteristic information is (1,0.98); the first symbol identification information is 0, and the second symbol identification information is 1; the third symbol identification information is 0, and the fourth symbol identification information is 0; and according to the analysis, the second associated dialogue data has neither question mark nor exclamation mark, and the fifth symbol identification information and the sixth symbol identification information are both 0, so that the target content characteristic information is (1,0.95,1,0.98,0,1,0,0,0,0).

In the above, it is described in detail how to obtain the target content feature information corresponding to the dialogue data to be analyzed according to the dialogue data to be analyzed and the associated dialogue data, and according to the above description, the dialogue processing method provided in the first embodiment of the present application does not need to identify according to keywords when determining whether the dialogue data to be analyzed is the dialogue data for consultation problems, but obtains the multidimensional content feature information corresponding to the dialogue data to be analyzed through the contextual dialogue data corresponding to the dialogue data to be analyzed, so as to increase the accuracy of the final decision result.

Hereinafter, a detailed description will be given of how to obtain target-related feature information corresponding to dialogue data to be analyzed.

When only the first associated dialogue data corresponding to the dialogue data to be analyzed is obtained in the above-described step S201, the target associated feature information corresponding to the dialogue data to be analyzed is obtained by: acquiring a first user type corresponding to a speaker of the dialogue data to be analyzed; judging whether the speaker of the dialogue data to be analyzed is the same as the speaker of the first associated dialogue data or not, and obtaining a first judgment result; acquiring a time interval between the dialogue data to be analyzed and the dialogue data to be analyzed of first associated dialogue data as first time interval data; and obtaining the target associated feature information according to the first user type, the first judging result and the first time interval data.

For example, the dialogue data to be analyzed is "speaker C (15:30:00): you get a consultation that i have a question … …? "the first associated dialog data is" speaker B (10:10:00): good, if there are other problems, feedback to me "; if the user type is 1 as the client, the user type is 0 as the customer service, the speaker is not the same as 0, the speaker is the same as 1, and the unit of the first time interval data is seconds, the first user type is 1, the first judgment result is 0, the first time interval data is 19200 seconds, and thus the target associated feature information is (1, 0, 19200).

It should be noted that, if in the above step S201, the second associated dialogue data corresponding to the dialogue data to be analyzed is also obtained, the target associated feature information corresponding to the dialogue data to be analyzed is also obtained by the following method: acquiring a second user type corresponding to a speaker of the second associated dialogue data; judging whether the speaker of the dialogue data to be analyzed is the same as the speaker of the second associated dialogue data or not, and obtaining a second judgment result; acquiring a time interval between the dialogue data to be analyzed and the second associated dialogue data as second time interval data; and obtaining the target associated characteristic information according to the first user type, the first judging result, the first time interval data, the second user type, the second judging result and the second time interval data.

In the above, it is described in detail how, according to the dialogue data to be analyzed and the associated dialogue data, the obtained multidimensional feature information corresponding to the dialogue data to be analyzed, such as the target content feature information and the target associated feature information, after the multidimensional feature information is obtained, whether the dialogue data to be analyzed is used for the consultation problem can be determined according to the feature information.

After step S202, step S203 is executed to determine whether the dialogue data to be analyzed is used for a consultation problem according to the target content feature information and the target associated feature information.

The following is a description of how to determine whether the dialogue data to be analyzed is used for receiving the consultation problem according to the obtained target content feature information and target association feature information corresponding to the dialogue data to be analyzed.

In particular, a method of determining whether dialogue data to be analyzed is dialogue data for consultation problems, comprising: inputting the target content characteristic information and the target associated characteristic information into a problem decision model to obtain a target decision result, wherein the problem decision model is used for determining whether the dialogue data to be analyzed is used for consulting a problem or not according to the obtained target content characteristic information and the target associated characteristic information, and the target decision result is used for identifying whether the dialogue data to be analyzed is used for consulting the problem or not.

It should be noted that, the problem decision model used in the first embodiment of the present application is a random forest model obtained by training in advance, and the problem decision model is obtained by the following method: acquiring relevant dialogue data for consultation problems from historical dialogue data as positive sample dialogue data, and acquiring relevant dialogue data which is not used for consultation problems as negative sample dialogue data, wherein the preset correlation conditions are met between the dialogue data in the positive sample dialogue data and the dialogue data in the negative sample dialogue data; training to obtain the problem decision model by using the positive sample dialogue data and the negative sample data as sample data; the sample data comprises sample dialogue data to be analyzed and associated dialogue data meeting the preset correlation condition between the sample dialogue data to be analyzed and the problem decision model is used for determining whether the sample dialogue data to be analyzed is used for consulting a problem according to input sample content characteristic information and sample associated characteristic information corresponding to the sample dialogue data to be analyzed.

In order to increase accuracy of decision results of the problem decision model, the first recognition model, the second recognition model and the problem decision model used in the dialogue data processing method provided in the first embodiment of the present application may also be jointly trained to obtain the three models after convergence, and specifically, the process of the joint training includes: acquiring relevant dialogue data for consultation problems from historical dialogue data as positive sample dialogue data, and acquiring relevant dialogue data which is not used for consultation problems as negative sample dialogue data, wherein the preset correlation conditions are met between the dialogue data in the positive sample dialogue data and the dialogue data in the negative sample dialogue data; obtaining target sample dialogue data according to the positive sample dialogue data and the negative sample dialogue data, wherein the target sample dialogue data comprises sample dialogue data to be analyzed and associated sample dialogue data, and the associated sample dialogue data and the sample dialogue data to be analyzed meet the preset correlation condition; training the first recognition model by using the sample dialogue data to be analyzed, and training the second recognition model by using the associated sample dialogue data to obtain sample content characteristic information corresponding to the sample dialogue data to be analyzed; sample association characteristic information corresponding to the sample dialogue data to be analyzed is obtained; training the problem decision model by using the sample content characteristic information and the sample association characteristic information; in the training process, parameters of the first recognition model, the second recognition model and the problem decision model are adjusted by using a loss function which corresponds to the first recognition model, the second recognition model and the problem decision model together, so that the first recognition model, the second recognition model and the problem decision model reach preset convergence conditions.

After the problem decision model is obtained, a target decision result for identifying whether the dialogue data to be analyzed is the dialogue data for consulting the problem can be obtained according to the multidimensional feature information corresponding to the dialogue data to be analyzed.

For example, target content feature information and target associated feature information (1,0.95,1,0.98,0,1,0,0,0,0,1,0, 19200 …) are input into the problem decision model, and target decision results can be obtained.

As can be seen from the above description, according to the dialogue data processing method in the first embodiment of the present application, through obtaining, from original dialogue data, associated dialogue data satisfying a preset correlation condition with dialogue data to be analyzed, further obtaining multidimensional feature information corresponding to the dialogue data to be analyzed, and determining whether the dialogue data to be analyzed is used for a consultation problem according to the feature information; the method is more convenient and accurate than using manual or keyword-based identification.

The method further comprises the steps of: and if the dialogue data to be analyzed is determined to be used for consulting the problem, generating task data which corresponds to the dialogue data to be analyzed and is used for processing the problem.

As shown in fig. 1-B, when the method is applied to a server, the generating task data corresponding to the dialogue data to be analyzed for processing the problem includes: providing the dialogue data to be analyzed to a client; receiving a data request message sent by a client for generating the task data; and generating the task data according to the data request message. And when the method is applied to a client, the generating task data corresponding to the dialogue data to be analyzed for processing the problem comprises the following steps: receiving the dialogue data to be analyzed provided by a server; generating a data request message for generating the task data according to the dialogue data to be analyzed; and sending the data request message to the server.

That is, when it is determined that the dialogue data to be analyzed is used for the consultation problem through the dialogue data processing method, corresponding task data, for example, a task sheet, can be generated and pushed to related persons, such as client persons and corresponding background technicians.

For example, as shown in fig. 1-B and fig. 1-C, after the client obtains the original session data, the original session data may be sent to the server; after the server side obtains the original dialogue data, the dialogue data processing method shown in the first embodiment of the application is used for obtaining dialogue data for consulting a problem contained in the original dialogue data, and pushing the dialogue data for consulting the problem to the client side in a problem list mode; after obtaining the problem list, the client can display the problem list, and can confirm whether to generate task data which corresponds to dialogue data for consulting the problem in the problem list and is used for processing the problem contained in the problem list to the user; if the user confirms that the task data can be generated, the client sends a data request message for obtaining target task data to the server, and after obtaining the data request message, the server can generate target task data corresponding to the problem in the problem list and push the target task data to the client; after obtaining the task data, the client can display the task data for the user to confirm.

In summary, the method for processing dialogue data provided in the first embodiment of the present application includes: acquiring dialogue data to be analyzed, and acquiring associated dialogue data corresponding to the dialogue data to be analyzed, wherein a preset correlation condition is met between the associated dialogue data and the dialogue data to be analyzed; according to the dialogue data to be analyzed and the associated dialogue data, obtaining target content characteristic information and target associated characteristic information corresponding to the dialogue data to be analyzed; and determining whether the dialogue data to be analyzed is used for consulting a problem according to the target content characteristic information and the target associated characteristic information. According to the method, whether the dialogue data are used for consulting the problem is not needed in a manual mode or according to the keyword information, and the dialogue data to be analyzed and the associated dialogue data meeting the preset correlation condition between the dialogue data to be analyzed are obtained, the multidimensional characteristic information corresponding to the dialogue data to be analyzed in the dialogue process is obtained, and whether the dialogue data to be analyzed are used for consulting the problem is determined according to the multidimensional characteristic information, so that the dialogue data related to the problem can be conveniently and accurately obtained.

The second embodiment of the present application also provides another session data processing method, which is a specific scenario processing method, please refer to fig. 3, which is a flowchart of another session data processing method provided in the second embodiment of the present application, wherein some of the steps are described in detail in the first embodiment of the present application, so that the description herein is relatively simple, and the relevant points are only referred to the partial description in the first embodiment of the present application, and the processing procedures described below are only illustrative.

Step S301, obtaining original dialogue data from a multi-user dialogue group for providing a target product consultation service to a user.

Step S302, obtaining current dialogue data from the original dialogue data as dialogue data to be analyzed, and obtaining associated dialogue data corresponding to the dialogue data to be analyzed, wherein a preset correlation condition is satisfied between the associated dialogue data and the dialogue data to be analyzed.

Step S303, obtaining target content characteristic information and target associated characteristic information corresponding to the dialogue data to be analyzed according to the dialogue data to be analyzed and the associated dialogue data.

Step S304, determining whether the dialogue data to be analyzed is dialogue data for consulting the use problem of the target product according to the target content feature information and the target associated feature information.

In correspondence to the dialogue data processing method provided in the first embodiment of the present application, the third embodiment of the present application further provides a method for obtaining a problem decision model, please refer to fig. 4, which is a flowchart of a method for obtaining a problem decision model provided in the third embodiment of the present application, wherein some steps have been described in detail in the first embodiment of the present application, so the description herein is relatively simple, and the relevant points are only referred to in some descriptions in the first embodiment of the present application, and the processing procedures described below are only illustrative.

Step S401, acquiring relevant dialogue data for consulting a problem from historical dialogue data as positive sample dialogue data, and acquiring relevant dialogue data not for consulting a problem as negative sample dialogue data, wherein the preset correlation condition is satisfied between dialogue data in the positive sample dialogue data and dialogue data in the negative sample dialogue data.

Step S402, training to obtain a problem decision model by using the positive sample dialogue data and the negative sample data as sample data; the sample data comprises sample dialogue data to be analyzed and associated dialogue data meeting the preset correlation condition between the sample dialogue data to be analyzed and the problem decision model is used for determining whether the sample dialogue data to be analyzed is used for consulting a problem according to input sample content characteristic information and sample associated characteristic information corresponding to the sample dialogue data to be analyzed.

In correspondence with a session data processing method provided in the first embodiment of the present application, a session data processing device is further provided in the fourth embodiment of the present application, please refer to fig. 5, which is a schematic diagram of a session data processing device provided in the fourth embodiment of the present application, and since the device embodiment is substantially similar to the method embodiment, the description is relatively simple, and the relevant point is referred to the part of the description of the method embodiment, and the device embodiment described below is merely illustrative. A dialogue data processing device provided in a fourth embodiment of the present application includes the following parts:

a data obtaining unit 501, configured to obtain dialogue data to be analyzed, and obtain associated dialogue data corresponding to the dialogue data to be analyzed, where a preset correlation condition is satisfied between the associated dialogue data and the dialogue data to be analyzed.

And the feature information obtaining unit 502 is configured to obtain, according to the dialogue data to be analyzed and the associated dialogue data, target content feature information and target associated feature information corresponding to the dialogue data to be analyzed.

A determining unit 503, configured to determine whether the dialogue data to be analyzed is used for a consultation problem according to the target content feature information and the target associated feature information.

In correspondence with the method for processing dialogue data provided in the first embodiment of the present application, the fifth embodiment of the present application further provides an electronic device, please refer to fig. 6, which is a schematic diagram of an electronic device provided in the fifth embodiment of the present application, and since the electronic device embodiment is substantially similar to the method embodiment, the description is relatively simple, and the relevant point is referred to the partial description of the method embodiment, and the electronic device embodiment described below is merely illustrative. An electronic device provided in a fifth embodiment of the present application includes:

a processor 601;

a memory 602 for storing a program for a dialogue data processing method, the device, after powering on and running the program for the dialogue data processing method by the processor, performs the steps of:

In correspondence with another session data processing method provided in the second embodiment of the present application, the sixth embodiment of the present application further provides another session data processing device, please refer to fig. 7, which is a schematic diagram of another session data processing device provided in the sixth embodiment of the present application, and since the device embodiment is substantially similar to the method embodiment, the description is relatively simple, and the relevant point is referred to the part of the description of the method embodiment, and the device embodiment described below is merely illustrative. A dialogue data processing device provided in a sixth embodiment of the present application includes the following parts:

the original dialogue data acquisition unit 701 acquires original dialogue data from a multi-user dialogue group for providing a target product consultation service to a user.

The data obtaining unit 702 obtains current dialogue data from the original dialogue data as dialogue data to be analyzed, and obtains associated dialogue data corresponding to the dialogue data to be analyzed, wherein a preset correlation condition is satisfied between the associated dialogue data and the dialogue data to be analyzed.

A feature information obtaining unit 703, configured to obtain target content feature information and target associated feature information corresponding to the dialogue data to be analyzed according to the dialogue data to be analyzed and the associated dialogue data.

And a determining unit 704, configured to determine whether the dialogue data to be analyzed is dialogue data for consulting a use problem of the target product according to the target content feature information and the target associated feature information.

In correspondence with another dialogue data processing method provided in the second embodiment of the present application, the seventh embodiment of the present application further provides an electronic device, and since the electronic device embodiment is substantially similar to the method embodiment, the description is relatively simple, and the relevant points are merely referred to in the partial description of the method embodiment, and the electronic device embodiment described below is merely illustrative. An electronic device provided in a seventh embodiment of the present application includes:

a processor;

In correspondence with the method for obtaining the problem decision model provided in the third embodiment of the present application, the eighth embodiment of the present application further provides an apparatus for obtaining the problem decision model, please refer to fig. 8, which is a schematic diagram of an apparatus for obtaining the problem decision model provided in the eighth embodiment of the present application, and since the apparatus embodiment is substantially similar to the method embodiment, the description is relatively simple, and the relevant points refer to a part of the description of the method embodiment, and the apparatus embodiment described below is merely schematic. An obtaining device of a problem decision model provided in an eighth embodiment of the present application includes the following parts:

A sample data obtaining unit 801, configured to obtain relevant dialogue data for consulting a problem from historical dialogue data as positive sample dialogue data, and obtain relevant dialogue data not for consulting a problem as negative sample dialogue data, where the preset correlation condition is satisfied between dialogue data in the positive sample dialogue data and dialogue data in the negative sample dialogue data.

A training unit 802, configured to use the positive sample dialogue data and the negative sample data as sample data, and train to obtain a problem decision model; the sample data comprises sample dialogue data to be analyzed and associated dialogue data meeting the preset correlation condition between the sample dialogue data to be analyzed and the problem decision model is used for determining whether the sample dialogue data to be analyzed is used for consulting a problem according to input sample content characteristic information and sample associated characteristic information corresponding to the sample dialogue data to be analyzed.

Corresponding to the method for obtaining a problem decision model provided in the third embodiment of the present application, the ninth embodiment of the present application further provides an electronic device, and since the electronic device embodiment is substantially similar to the method embodiment, the description is relatively simple, and the relevant points are only referred to in the partial description of the method embodiment, and the electronic device embodiment described below is merely illustrative. An electronic device provided in a ninth embodiment of the present application includes:

A processor;

In correspondence to the session data processing method provided in the first embodiment of the present application, the tenth embodiment of the present application further provides a data pushing method, which is a specific scenario processing method, please refer to fig. 9, which is a flowchart of a data pushing method provided in the tenth embodiment of the present application, wherein part of the steps are described in detail in the first embodiment of the present application, so that the description herein is relatively simple, and the relevant points refer to part of the description in the first embodiment of the present application, and the processing procedures described below are only illustrative.

Step S901, obtain the original dialogue data sent by the client.

Step S902, obtaining target problem list data corresponding to the original dialogue data, where the target problem list data is data including at least one piece of dialogue data for consulting a problem obtained by using the dialogue data processing method provided in the first embodiment of the present application.

Step S903, pushing the target problem list data to the client.

Corresponding to the session data processing method provided in the first embodiment and the data pushing method provided in the tenth embodiment of the present application, the eleventh embodiment of the present application further provides a data display method, which is a specific scenario processing method, please refer to fig. 10, which is a flowchart of a data display method provided in the eleventh embodiment of the present application, wherein part of the steps are described in detail in the foregoing embodiments, so that the description herein is relatively simple, and the relevant points refer to part of the description in the foregoing embodiments, and the processing procedure described below is only illustrative.

In step S1001, original dialogue data is acquired.

Step S1002, sending the original dialogue data to a server.

Step S1003, receiving target problem list data pushed by the server and corresponding to the original dialogue data, where the target problem list data is data, obtained by the server by using the dialogue data processing method provided by the first embodiment of the present application, including at least one piece of dialogue data for consulting a problem.

Step S1004, displaying the target problem list data.

In correspondence with a data pushing method provided in the tenth embodiment of the present application, a data pushing device is further provided in the twelfth embodiment of the present application, please refer to fig. 11, which is a schematic diagram of a data pushing device provided in the twelfth embodiment of the present application, and since the device embodiment is substantially similar to the method embodiment, the description is relatively simple, and the relevant points refer to a part of the description of the method embodiment, and the device embodiment described below is merely illustrative. A data estimation device provided in a twelfth embodiment of the present application includes:

The data acquisition unit 1101 is configured to acquire original dialogue data sent by a client.

The question list data obtaining unit 1102 is configured to obtain target question list data corresponding to the original dialogue data, where the target question list data is data obtained by the dialogue data processing method provided in the first embodiment of the present application and includes at least one piece of dialogue data for consulting a question.

And a pushing unit 1103, configured to push the target problem list data to the client.

Optionally, the apparatus further includes: a request message obtaining unit, configured to obtain a data request message sent by the client for obtaining target task data, where the target task data is data for processing a problem corresponding to dialogue data for consulting the problem in the target problem list data; a task data generating unit, configured to generate the target task data according to the target problem list data; and the task data display unit is used for pushing the target task data to the client.

Corresponding to a data pushing method provided in the tenth embodiment of the present application, the thirteenth embodiment of the present application further provides an electronic device, and since the electronic device embodiment is substantially similar to the method embodiment, the description is relatively simple, and the relevant points only need to be referred to in the partial description of the method embodiment, and the electronic device embodiment described below is merely illustrative. An electronic device provided in a thirteenth embodiment of the present application includes:

A processor;

a memory for storing a program of a data pushing method, the apparatus being powered on and executing the program of the data pushing method by the processor, and performing the steps of:

acquiring original dialogue data sent by a client;

acquiring target problem list data corresponding to the original dialogue data, wherein the target problem list data is data which is obtained by using the dialogue data processing method provided by the first embodiment of the application and comprises at least one piece of dialogue data for consulting a problem;

and pushing the target problem list data to the client.

In correspondence with a data display method provided in the eleventh embodiment of the present application, a data display device is further provided in the fourteenth embodiment of the present application, please refer to fig. 12, which is a schematic diagram of a data display device provided in the fourteenth embodiment of the present application, and since the device embodiments are substantially similar to the method embodiments, the description is relatively simple, and the relevant points refer to a part of the description of the method embodiments, and the device embodiments described below are only schematic. A data display device provided in a fourteenth embodiment of the present application includes the following parts:

A data acquisition unit 1201 for acquiring original dialogue data.

A data sending unit 1202, configured to send the original session data to a server.

The problem list data receiving unit 1203 is configured to receive target problem list data corresponding to the original dialogue data, where the target problem list data is data that is obtained by the server and includes at least one piece of dialogue data for consulting a problem by using the dialogue data processing method provided in the first embodiment of the present application.

And the display unit 1204 is used for displaying the target problem list data.

Optionally, the apparatus further includes: a request message sending unit, configured to send a data request message for obtaining target task data to the server, where the target task data is data for processing a problem corresponding to dialogue data for consulting the problem in the target problem list data; the task data receiving unit is used for receiving the target task data pushed by the server; and the task data display unit is used for displaying the target task data.

Corresponding to a data presentation method provided in the eleventh embodiment of the present application, the fifteenth embodiment of the present application further provides an electronic device, and since the electronic device embodiment is substantially similar to the method embodiment, the description is relatively simple, and the relevant points are only referred to in the partial description of the method embodiment, and the electronic device embodiment described below is merely illustrative. An electronic device provided in a fifteenth embodiment of the present application includes:

A processor;

a memory for storing a program of a data presentation method, the apparatus being powered on and executing the program of the data presentation method by the processor, and performing the steps of:

acquiring original dialogue data;

sending the original dialogue data to a server;

receiving target problem list data pushed by the server and corresponding to the original dialogue data, wherein the target problem list data is data which is obtained by the server through the dialogue data processing method provided by the first embodiment of the application and contains at least one piece of dialogue data for consulting a problem;

and displaying the target problem list data.

While the preferred embodiment has been described, it is not intended to limit the invention thereto, and any person skilled in the art may make variations and modifications without departing from the spirit and scope of the present invention, so that the scope of the present invention shall be defined by the claims of the present application.

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash memory (flashRAM). Memory is an example of computer-readable media.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer readable media, as defined herein, does not include non-transitory computer readable media (transmission media), such as modulated data signals and carrier waves.

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.

Claims

1. A method of processing dialogue data, comprising:

according to the dialogue data to be analyzed and the associated dialogue data, target content characteristic information and target associated characteristic information corresponding to the dialogue data to be analyzed are obtained, wherein the target content characteristic information corresponding to the dialogue data to be analyzed is obtained through the following method:

inputting the dialogue data to be analyzed into a first recognition model, and acquiring first content characteristic information corresponding to the dialogue data to be analyzed, wherein the first recognition model is used for recognizing whether the dialogue data to be analyzed is used for consulting a first problem; inputting first associated dialogue data corresponding to the dialogue data to be analyzed into a second recognition model, and obtaining second content characteristic information corresponding to the second dialogue data, wherein the second recognition model is used for recognizing whether the first associated dialogue data is dialogue data for ending a dialogue corresponding to consultation of a second problem; obtaining the target content characteristic information according to the first content characteristic information and the second content characteristic information;

The target associated characteristic information corresponding to the dialogue data to be analyzed is obtained by the following method:

acquiring a first user type corresponding to a speaker of the dialogue data to be analyzed; judging whether the speaker of the dialogue data to be analyzed is the same as the speaker of the first associated dialogue data or not, and obtaining a first judgment result; acquiring a time interval between the dialogue data to be analyzed and the dialogue data to be analyzed of first associated dialogue data as first time interval data; obtaining the target associated feature information according to the first user type, the first judgment result and the first time interval data;

2. The dialogue data processing method according to claim 1, characterized by further comprising: acquiring original dialogue data;

the step of acquiring dialogue data to be analyzed and acquiring associated dialogue data corresponding to the dialogue data to be analyzed comprises the following steps:

and acquiring the dialogue data to be analyzed from the original dialogue data, and acquiring the associated dialogue data corresponding to the dialogue data to be analyzed.

3. The dialogue data processing method according to claim 2, wherein the acquiring the dialogue data to be analyzed from the original dialogue data, and the acquiring the associated dialogue data corresponding to the dialogue data to be analyzed, comprises:

4. The method for processing dialogue data according to claim 1, wherein said determining whether the dialogue data to be analyzed is used for a consultation problem according to the target content feature information and the target associated feature information comprises:

5. The dialogue data processing method according to claim 2, wherein the acquiring the dialogue data to be analyzed and the acquiring the associated dialogue data corresponding to the dialogue data to be analyzed from the original dialogue data further comprises:

6. The dialogue data processing method according to claim 1, wherein the target content feature information corresponding to the dialogue data to be analyzed is further obtained by:

7. The dialogue data processing method according to claim 1, wherein the target association characteristic information corresponding to the dialogue data to be analyzed is further obtained by:

Acquiring a second user type corresponding to a speaker of second associated dialogue data;

8. The dialogue data processing method according to claim 1, characterized by further comprising:

and if the dialogue data to be analyzed is determined to be used for consulting the problem, generating task data which corresponds to the dialogue data to be analyzed and is used for processing the problem.

9. The dialogue data processing method according to claim 8, wherein the method is applied to a server side, and the generating task data for processing the problem corresponding to the dialogue data to be analyzed includes:

Providing the dialogue data to be analyzed to a client;

receiving a data request message sent by a client for generating the task data;

and generating the task data according to the data request message.

10. The dialogue data processing method according to claim 8, wherein the method is applied to a client, the generating task data for processing the problem corresponding to the dialogue data to be analyzed includes:

receiving the dialogue data to be analyzed provided by a server;

and sending the data request message to the server.

11. The dialogue data processing method of claim 4, wherein the problem decision model is obtained by:

Training to obtain the problem decision model by using the positive sample dialogue data and the negative sample dialogue data as sample data;

12. The dialogue data processing method according to claim 1, wherein the obtaining the target content feature information from the first content feature information and the second content feature information includes:

13. The dialogue data processing method according to claim 1, wherein the first content feature information includes: the first result identification information is used for identifying whether the dialogue data to be analyzed corresponds to the first problem or not, and the first score information is used for indicating the accuracy degree of the first result identification information;

14. The method for processing dialogue data according to claim 6, wherein the acquiring third content feature information corresponding to the second associated dialogue data includes:

15. The dialog data processing method of claim 2, wherein the raw dialog data includes text data, image data, audio data, and video data.

16. A method of processing dialogue data, comprising:

17. The data pushing method is characterized by being applied to a server and comprising the following steps:

acquiring original dialogue data sent by a client;

obtaining target problem list data corresponding to the original dialogue data, wherein the target problem list data is data which is obtained by using the dialogue data processing method according to any one of claims 1-15 and comprises at least one piece of dialogue data for consulting a problem;

and pushing the target problem list data to the client.

18. The data pushing method of claim 17, further comprising:

acquiring a data request message sent by the client for acquiring target task data, wherein the target task data is data which corresponds to dialogue data for consulting a problem in the target problem list data and is used for processing the problem;

generating the target task data according to the target problem list data;

and pushing the target task data to the client.

19. The data display method is characterized by being applied to a client and comprising the following steps of:

acquiring original dialogue data;

sending the original dialogue data to a server;

receiving target problem list data pushed by the server and corresponding to the original dialogue data, wherein the target problem list data is data which is obtained by the server through any one of the dialogue data processing methods in claims 1-15 and contains at least one piece of dialogue data for consulting a problem;

and displaying the target problem list data.

20. The data presentation method of claim 19, further comprising:

sending a data request message for obtaining target task data to the server, wherein the target task data is data which corresponds to dialogue data for consulting a problem in the target problem list data and is used for processing the problem;

receiving the target task data pushed by the server;

and displaying the target task data.

21. A dialogue data processing apparatus, comprising:

the data acquisition unit is used for acquiring dialogue data to be analyzed from the original dialogue data and acquiring associated dialogue data corresponding to the dialogue data to be analyzed, wherein the associated dialogue data and the dialogue data to be analyzed meet preset correlation conditions;

The feature information acquisition unit is used for acquiring target content feature information and target associated feature information corresponding to the dialogue data to be analyzed according to the dialogue data to be analyzed and the associated dialogue data, wherein the target content feature information corresponding to the dialogue data to be analyzed is acquired through the following method:

22. An electronic device, comprising:

a processor;

23. A dialogue data processing apparatus, comprising:

24. An electronic device, comprising:

a processor;

25. The data pushing device is characterized by being applied to a server and comprising the following components:

the data acquisition unit is used for acquiring original dialogue data sent by the client;

a question list data obtaining unit configured to obtain target question list data corresponding to the original dialogue data, where the target question list data is data including at least one piece of dialogue data for consultation of a question obtained using the dialogue data processing method according to any one of claims 1 to 15;

And the pushing unit is used for pushing the target problem list data to the client.

26. An electronic device, applied to a server, comprising:

a processor;

acquiring original dialogue data sent by a client;

and pushing the target problem list data to the client.

27. A data presentation device, for application to a client, comprising:

the data acquisition unit is used for acquiring original dialogue data;

the data sending unit is used for sending the original dialogue data to the server;

a question list data receiving unit, configured to receive target question list data corresponding to the original dialogue data, where the target question list data is data that is obtained by the server using any one of the dialogue data processing methods in claims 1-15 and includes at least one piece of dialogue data for consulting a question;

And the display unit is used for displaying the target problem list data.

28. An electronic device, for application to a client, comprising:

a processor;

acquiring original dialogue data;

sending the original dialogue data to a server;

and displaying the target problem list data.