Information Push Method, Device, Storage Medium, and Electronic Device

Abstract

The present application discloses an information push method, device, storage medium, and electronic device, which relate to the field of financial technology. The method includes: obtaining call voices generated during the process of pushing products to users; performing voice analysis on the call voices through a language analysis model to obtain voice information corresponding to the call voices, where the voice information includes semantic information and emotional information; determining push information corresponding to the call voices according to the semantic information and the emotional information to obtain target push information; and sending the target push information to the users. Through the present application, the problem in the related art that it is difficult to accurately push information according to diverse reply information of different users during product pushing is solved.

Description

Technical Field

[0001] This application relates to the field of fintech, and in particular, to an information push method, device, storage medium, and electronic device. Background Technique

[0002] Currently, in the process of product promotion, the words and techniques used by promoters often come from their own training or learning from previous promotion cases in materials, so as to give users good feedback when promoting products to different users, and thus successfully complete product promotion.

[0003] However, for promoters without rich experience, due to the lack of experience and ability, they will not know how to give accurate and satisfactory responses when promoting products. Therefore, some simple basic response words are usually provided to promoters to improve the accuracy of the response words. However, since most of the current response words are template-based and rigid, they cannot change the response words in real time and objectively according to the user's reaction, resulting in a wrong judgment of the user's psychological needs, leading to the failure of promotion and even possible user disgust.

[0004] In view of the problem that it is difficult to accurately push information according to the diverse response information of different users when pushing products in the related art, no effective solution has been proposed yet. Summary of the Invention

[0005] This application provides an information push method, device, storage medium, and electronic device to solve the problem that it is difficult to accurately push information according to the diverse response information of different users when pushing products in the related art.

[0006] According to one aspect of this application, an information push method is provided. The method includes: obtaining the call voice generated during the process of product push to the user; performing voice analysis on the call voice through a language analysis model to obtain the voice information corresponding to the call voice, where the voice information includes semantic information and emotional information; determining the push information corresponding to the call voice according to the semantic information and emotional information to obtain the target push information; and sending the target push information to the user.

[0007] Optionally, the language analysis model includes a speech recognition sub-model. Performing voice analysis on the call voice through the language analysis model to obtain the voice information corresponding to the call voice includes: performing semantic recognition on the call voice through the speech recognition sub-model to obtain semantic information; performing emotional analysis on the semantic information to obtain emotional information; and combining the semantic information and the emotional information to obtain the voice information.

[0008] Optionally, the semantic information includes content information and voice information. The language analysis model further includes a dictionary library and a sentiment analysis sub-model. Sentiment analysis is performed on the semantic information to obtain sentiment information, including: determining multiple keywords from the content information, matching the multiple keywords with sample keywords in the dictionary library to obtain the first sentiment information corresponding to each keyword, where the dictionary library includes multiple sample keywords and the sample first sentiment information corresponding to each sample keyword; inputting the multiple keywords and voice information into the sentiment analysis sub-model to process and obtain the second sentiment information, where the sentiment analysis sub-model is trained by sample keywords, sample voice information, and sample second sentiment information; combining the first sentiment information and the second sentiment information to obtain the sentiment information.

[0009] Optionally, before inputting the multiple keywords and voice information into the sentiment analysis sub-model to process and obtain the second sentiment information, the method further includes: obtaining multiple sample keywords from the dictionary library, and determining the sample voice information corresponding to each sample keyword and the sample second sentiment information corresponding to each sample keyword; determining each sample keyword, the sample voice information corresponding to the sample keyword, and the sample second sentiment information corresponding to the sample keyword as a set of first sample data to obtain multiple sets of first sample data; training a first machine learning model with the multiple sets of first sample data to obtain the sentiment analysis sub-model.

[0010] Optionally, the semantic information further includes scene information and the relationship information between multiple statements. According to the semantic information and the sentiment information, the push information corresponding to the call voice is determined to obtain the target push information, including: inputting the semantic information and the sentiment information into the target model to obtain multiple push information corresponding to the voice information, where the target model is trained by sample voice information and sample push information; determining the target push information from the multiple push information according to the scene information and the relationship information between multiple statements.

[0011] Optionally, before inputting the semantic information and the sentiment information into the target model to obtain multiple push information corresponding to the voice information, the method further includes: determining each historical voice information and the historical push information corresponding to the historical voice information as a set of second sample data to obtain multiple sets of second sample data; training a second machine learning model with the multiple sets of second sample data to obtain the target model.

[0012] Optionally, after sending the target push information to the user, the method further includes: obtaining the feedback information of the user for the target push information; determining the score corresponding to the feedback information and comparing the score with the threshold; in the case where the score is higher than the threshold, obtaining the target voice information corresponding to the target push information and adding the target push information and the target voice information to the second sample data.

[0013] According to another aspect of the present application, an information push device is provided. The device includes: a first acquisition unit for acquiring call voices generated during the process of pushing products to users; an analysis unit for performing voice analysis on the call voices through a language analysis model to obtain voice information corresponding to the call voices, where the voice information includes semantic information and emotional information; a first determination unit for determining push information corresponding to the call voices according to the semantic information and the emotional information to obtain target push information; and a sending unit for sending the target push information to the users.

[0014] According to another aspect of the embodiments of the present invention, a computer storage medium is further provided. The computer storage medium is used to store a program, where when the program runs, it controls a device where the computer storage medium is located to execute an information push method.

[0015] According to another aspect of the embodiments of the present invention, an electronic device is further provided, including one or more processors and a memory; computer-readable instructions are stored in the memory, and the processors are used to run the computer-readable instructions, where when the computer-readable instructions run, they execute an information push method.

[0016] Through the present application, the following steps are adopted: acquiring call voices generated during the process of pushing products to users; performing voice analysis on the call voices through a language analysis model to obtain voice information corresponding to the call voices, where the voice information includes semantic information and emotional information; determining push information corresponding to the call voices according to the semantic information and the emotional information to obtain target push information; and sending the target push information to the users. This solves the problem in the related art that it is difficult to accurately push information according to the diverse reply information of different users during product pushing. By acquiring the voice information during communication with users, obtaining semantic information and emotional information from the voice information according to the language analysis model, and determining the most suitable reply information from multiple push information according to the semantic information and the emotional information, the effect of accurately determining the push information corresponding to the target user is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0018] Figure 1 is a flowchart of the information push method provided by the embodiments of the present application;

[0019] Figure 2 is a schematic diagram of the information push device provided by the embodiments of the present application;

[0020] Figure 3It is a schematic diagram of an electronic device provided according to an embodiment of the present application. Detailed implementation manners

[0021] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments may be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0022] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data used may be interchanged under appropriate circumstances so as to describe the embodiments of the present application here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0024] It should be noted that the relevant information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in the present disclosure are all information and data authorized by the user or fully authorized by all parties. For example, an interface is provided between the present system and relevant users or institutions. Before obtaining relevant information, a request for obtaining information needs to be sent to the aforementioned users or institutions through the interface, and after receiving the consent information feedback from the aforementioned users or institutions, the relevant information is obtained.

[0025] It should be noted that the information push method, device, storage medium, and electronic device determined in the present disclosure can be used in the field of fintech, and can also be used in any field other than the field of fintech. The application fields of the information push method, device, storage medium, and electronic device determined in the present disclosure are not limited.

[0026] For the convenience of description, some nouns or terms involved in the embodiments of the present application are described below:

[0027] NLP model: Natural Language Processing natural language processing model.

[0028] According to an embodiment of the present application, an information push method is provided.

[0029] Figure 1 It is a flowchart of the information push method according to an embodiment of the present application. As Figure 1 shown, the method includes the following steps:

[0030] Step S101, obtain the call voice generated during the process of pushing products to users.

[0031] Specifically, the way of product push can be that a promoter makes a phone call to promote to the user, or an artificial intelligence voice promotes the product to the user. During the promotion, the call voice of the user can be obtained, so as to determine the user's attitude towards this promotion and the feedback content through the call voice.

[0032] For example, the promotion content can be: "Do you know about Product A?", and the user's call voice can be: "I have tried this product".

[0033] Step S102, perform voice analysis on the call voice through a language analysis model to obtain the voice information corresponding to the call voice, where the voice information includes semantic information and emotional information.

[0034] Specifically, after obtaining the call voice, the language analysis model can analyze the call voice from the semantic dimension and the emotional dimension, so as to judge the text information and the user's attitude information conveyed by the voice information.

[0035] For example, the user's call voice is: "I haven't heard of it. Don't harass me anymore in the future". At this time, the user's text information can be obtained: "haven't heard of it, don't, harass", and at the same time, emotional judgment is made according to the user's tone, so as to determine the user's reaction to this promotion from two dimensions at the same time, and determine the corresponding words according to this reaction.

[0036] In order to accurately identify the call voice and determine the voice information in the user's voice according to the call voice. Optionally, in the information push method provided by the embodiment of the present application, the language analysis model includes a speech recognition sub-model. Performing voice analysis on the call voice through the language analysis model to obtain the voice information corresponding to the call voice includes: performing semantic recognition on the call voice through the speech recognition sub-model to obtain semantic information; performing emotional analysis on the semantic information to obtain emotional information; combining the semantic information and the emotional information to obtain voice information.

[0037] Specifically, the speech recognition sub-model can be speech recognition software, which is responsible for recognizing the collected customer speech information and converting it into corresponding semantic information. Among them, the semantic information can include text information and voice information. That is, the intonation is recognized through the acoustic model in the speech recognition software, and the text is recognized according to the language model, so as to disassemble and classify the speech information.

[0038] Furthermore, after obtaining the semantic information, the attitude and thoughts of the user towards this promotion expressed in the customer speech information, that is, the emotional information, can be comprehensively analyzed based on different words in the semantic information, the tones corresponding to different words, and the intonation of the whole sentence.

[0039] After obtaining the semantic information and the emotional information, the two kinds of information can be summarized to obtain the complete speech information, and the push information can be obtained according to the speech information. In this embodiment, by determining the semantic information and the emotional information, and determining the speech information according to the semantic information and the emotional information, the effect of improving the accuracy of the push information is achieved.

[0040] In order to accurately determine the emotional information, optionally, in the information push method provided in the embodiment of the present application, the semantic information includes content information and voice information, and the language analysis model further includes a dictionary library and an emotion analysis sub-model. The emotion analysis of the semantic information to obtain the emotional information includes: determining multiple keywords from the content information, matching the multiple keywords with the sample keywords in the dictionary library to obtain the first emotional information corresponding to each keyword, where the dictionary library includes multiple sample keywords and the sample first emotional information corresponding to each sample keyword; inputting the multiple keywords and the voice information into the emotion analysis sub-model, and processing to obtain the second emotional information, where the emotion analysis sub-model is trained by the sample keywords, the sample voice information, and the sample second emotional information; combining the first emotional information and the second emotional information to obtain the emotional information.

[0041] Specifically, the first emotional information can be the emotional information of words. The dictionary library can store multiple sample keywords and the sample first emotional information corresponding to each sample keyword. Among them, the sample keywords can be a large number of words for which the emotional information has been determined. For example, the emotional information of "continue talking" can be happy, and the emotional information of "stop talking" can be bored.

[0042] After obtaining the semantic information of the call voice, the content information in the semantic information can be segmented first to obtain multiple keywords, and the first emotional information of each keyword can be marked according to the emotional information of the words in the dictionary library, so as to determine the first emotional information corresponding to the semantic information.

[0043] Further, after word segmentation, each keyword and its corresponding voice information can be input into the sentiment analysis sub-model. The sentiment analysis sub-model analyzes the keyword and its voice information to determine the second sentiment information corresponding to each keyword. The sentiment analysis sub-model can be an NLP model, and the second sentiment information can be the sentiment information obtained by analyzing each keyword in combination with the voice information. For example, the user's content information can be "I'm busy now, stop talking". At this time, if only the sentiment information recorded in the dictionary is analyzed, it will be judged that the user's sentiment is boredom. However, if the user's tone is calm at this time, it indicates that the user is doing other things and is not bored with this promotion. Therefore, by analyzing the words and tones simultaneously through the sentiment analysis sub-model, the user's true sentiment can be determined more accurately.

[0044] After obtaining the first sentiment information and the second sentiment information, the two types of information can be combined, and weights can be added. For example, the reference degree of the first sentiment information is 40%, and the reference degree of the second sentiment information is 60%, so as to accurately determine the information and sentiment expressed by the user from the user's call voice. This embodiment accurately determines the user's sentiment information from two dimensions, laying a foundation for accurately determining the user's attitude and tendency towards this promotion.

[0045] To improve the accuracy of the sentiment analysis sub-model in determining the second sentiment information, optionally, in the information push method provided in the embodiment of the present application, before inputting multiple keywords and voice information into the sentiment analysis sub-model to process and obtain the second sentiment information, the method further includes: obtaining multiple sample keywords from the dictionary, and determining the sample voice information corresponding to each sample keyword and the sample second sentiment information corresponding to each sample keyword; determining each sample keyword, the sample voice information corresponding to the sample keyword, and the sample second sentiment information corresponding to the sample keyword as a set of first sample data to obtain multiple sets of first sample data; training the first machine learning model with multiple sets of first sample data to obtain the sentiment analysis sub-model.

[0046] It should be noted that before using the sentiment analysis sub-model to determine the second sentiment information, it is necessary to first train the sentiment analysis sub-model based on the first sample data to improve the accuracy of judgment. Specifically, the first sample data may include sample keywords, sample voice information corresponding to each sample keyword, and sample second sentiment information corresponding to each sample keyword. The first sample data is input into the untrained sentiment analysis sub-model for training, and the training is completed when the sentiment analysis sub-model can correctly determine the sample second sentiment information corresponding to the sample keyword through the sample keyword and the sample voice information corresponding to each sample keyword, thereby obtaining an optimized sentiment analysis sub-model. In this embodiment, by optimizing the sentiment analysis sub-model, the accuracy of determining sentiment information is improved.

[0047] Step S103: Determine the push information corresponding to the call voice according to the semantic information and the sentiment information to obtain the target push information.

[0048] Specifically, after obtaining the semantic information and the sentiment information, the push information with the highest correlation degree with the semantic information and the sentiment information can be determined from a large number of promotion information according to the semantic information and the sentiment information, and the push information is determined as the target push information.

[0049] To accurately obtain the target push information, optionally, in the information push method provided in the embodiment of the present application, the semantic information further includes scene information and relationship information between multiple statements. Determining the push information corresponding to the call voice according to the semantic information and the sentiment information to obtain the target push information includes: inputting the semantic information and the sentiment information into the target model to obtain multiple push information corresponding to the voice information, where the target model is trained by sample voice information and sample push information; determining the target push information from the multiple push information according to the scene information and the relationship information between multiple statements.

[0050] Specifically, the semantic information and the sentiment information can be input into the target model, and the target model determines multiple push information corresponding to the voice information. The target model can perform big data analysis on the semantic information and the sentiment information and perform matching in the big data to obtain multiple push information with a relatively high correlation degree.

[0051] Furthermore, after obtaining the multiple push information, the scene information corresponding to the scene where the user is located and the context association information generated by multiple groups of conversations with the user can be combined to select the push information closest to the call voice from the multiple push information, thereby obtaining the target push information.

[0052] Optionally, in the information push method provided in the embodiments of the present application, before inputting the semantic information and the emotional information into the target model to obtain multiple push messages corresponding to the voice message, the method further includes: determining each historical voice message and the historical push message corresponding to the historical voice message as a set of second sample data, and obtaining multiple sets of second sample data; training the second machine learning model with the multiple sets of second sample data to obtain the target model.

[0053] It should be noted that before using the target model to determine multiple push messages, it is also necessary to first train the target model according to the second sample data, so as to improve the accuracy of judgment. Specifically, the second sample data may include historical voice messages and the historical push messages corresponding to each historical voice message. The second sample data is input into the untrained target model for training, and the training is completed when the target model can correctly obtain the correct historical push message through the historical voice message, so as to obtain the optimized target model. This embodiment improves the accuracy of determining the push message by optimizing the target model.

[0054] Step S104, sending the target push message to the user.

[0055] Specifically, after determining the target push message from the multiple push messages, the target push message can be directly sent to the user to complete the reply to the user, or it can be pushed to the mobile device of the promoter through the message push technology, and the target push message is displayed to the promoter through the display device, so that the promoter can reply according to the target push message.

[0056] The information push method provided in the embodiments of the present application obtains the call voice generated during the process of pushing products to the user; performs voice analysis on the call voice through a language analysis model to obtain the voice message corresponding to the call voice, where the voice message includes semantic information and emotional information; determines the push message corresponding to the call voice according to the semantic information and the emotional information to obtain the target push message; and sends the target push message to the user. This solves the problem in the related art that it is difficult to accurately push information in response to the diverse reply information of different users during product promotion. By obtaining the voice message during the communication with the user, obtaining the semantic information and the emotional information from the voice message according to the language analysis model, and determining the most suitable reply information from multiple push messages according to the semantic information and the emotional information, the effect of accurately determining the push message corresponding to the target user is achieved.

[0057] Optionally, in order to gradually improve the ability of the target model to determine push messages, in the information push method provided in the embodiments of the present application, after sending the target push information to the user, the method further includes: obtaining feedback information of the user on the target push information; determining a score corresponding to the feedback information, and comparing the score with a threshold; in the case where the score is higher than the threshold, obtaining target voice information corresponding to the target push information, and adding the target push information and the target voice information to the second sample data.

[0058] Specifically, after completing the promotion to the user, the information feedback by the user can be obtained, scored according to the feedback information, and it can be determined whether the target push message used in this promotion is accurate according to the score. In the case where the score is higher than the score threshold, the target push message pushed this time is determined as a valid message, and the target push message and the voice information corresponding to the target push message are added to the second sample data, so as to improve the second sample data and further train the target model, thereby achieving the effect of improving the push ability of the target model.

[0059] For example, the score can be manually calibrated through user evaluation. The user evaluation can be: excellent, then the score can correspond to 80 points, and the score threshold can be 60 points. Since the score is greater than the score threshold, the voice information and the corresponding target push information can be added to the second sample data, thereby completing the supplement of the sample data.

[0060] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0061] The embodiments of the present application also provide an information push device. It should be noted that the information push device of the embodiments of the present application can be used to execute the information push method provided by the embodiments of the present application. The following introduces the information push device provided by the embodiments of the present application.

[0062] Figure 2 is a schematic diagram of the information push device according to the embodiments of the present application. As Figure 2 shown, the device includes: a first acquisition unit 21, an analysis unit 22, a first determination unit 23, and a sending unit 24.

[0063] The first acquisition unit 21 is used to acquire the call voice generated during the process of pushing products to the user.

[0064] The analysis unit 22 is used to perform voice analysis on the call voice through a language analysis model to obtain voice information corresponding to the call voice, where the voice information includes semantic information and emotional information.

[0065] A first determination unit 23, configured to determine a push message corresponding to the call voice according to the semantic information and the emotion information, and obtain a target push message.

[0066] A sending unit 24, configured to send the target push message to the user.

[0067] The information push device provided by the embodiment of the present application obtains a call voice generated during the process of pushing a product to a user through a first acquisition unit 21; an analysis unit 22 performs voice analysis on the call voice through a language analysis model to obtain voice information corresponding to the call voice, where the voice information includes semantic information and emotion information. The first determination unit 23 determines a push message corresponding to the call voice according to the semantic information and the emotion information, and obtains a target push message; the sending unit 24 sends the target push message to the user; this solves the problem in the related art that it is difficult to accurately push information according to the diverse reply information of different users during product pushing. By obtaining the voice information during communication with the user, obtaining the semantic information and the emotion information from the voice information according to the language analysis model, and determining the most suitable reply information from multiple push messages according to the semantic information and the emotion information, the effect of accurately determining the push message corresponding to the target user is achieved.

[0068] Optionally, in the information push device provided by the embodiment of the present application, the language analysis model includes a speech recognition sub-model, and the analysis unit 22 includes: a recognition module, configured to perform semantic recognition on the call voice through the speech recognition sub-model to obtain semantic information; an analysis module, configured to perform emotion analysis on the semantic information to obtain emotion information; a combination module, configured to combine the semantic information and the emotion information to obtain voice information.

[0069] Optionally, in the information push device provided by the embodiment of the present application, the semantic information includes content information and voice information, the language analysis model further includes a dictionary library and an emotion analysis sub-model, and the analysis module includes: a determination sub-module, configured to determine multiple keywords from the content information, and match the multiple keywords with the sample keywords in the dictionary library to obtain the first emotion information corresponding to each keyword, where the dictionary library includes multiple sample keywords and the sample first emotion information corresponding to each sample keyword; a processing sub-module, configured to input the multiple keywords and the voice information into the emotion analysis sub-model for processing to obtain second emotion information, where the emotion analysis sub-model is trained by sample keywords, sample voice information, and sample second emotion information; a combination sub-module, configured to combine the first emotion information and the second emotion information to obtain emotion information.

[0070] Optionally, in the information push device provided in the embodiments of the present application, the device further includes: a second acquisition unit, configured to acquire a plurality of sample keywords from a dictionary library, and determine sample voice information corresponding to each sample keyword and sample second emotion information corresponding to each sample keyword; a second determination unit, configured to determine each sample keyword, the sample voice information corresponding to the sample keyword, and the sample second emotion information corresponding to the sample keyword as a set of first sample data, to obtain multiple sets of first sample data; a first training unit, configured to train a first machine learning model through the multiple sets of first sample data to obtain an emotion analysis sub-model.

[0071] Optionally, in the information push device provided in the embodiments of the present application, the semantic information further includes scenario information and relationship information between multiple statements. The first determination unit 23 includes: an input module, configured to input the semantic information and the emotion information into a target model to obtain multiple push messages corresponding to the voice information, where the target model is trained by sample voice information and sample push messages; a determination module, configured to determine a target push message from the multiple push messages according to the scenario information and the relationship information between the multiple statements.

[0072] Optionally, in the information push device provided in the embodiments of the present application, the device further includes: a third determination unit, configured to determine each historical voice information and the historical push message corresponding to the historical voice information as a set of second sample data, to obtain multiple sets of second sample data; a second training unit, configured to train a second machine learning model through the multiple sets of second sample data to obtain a target model.

[0073] Optionally, in the information push device provided in the embodiments of the present application, the device further includes: a third acquisition unit, configured to acquire feedback information of a user on a target push message; a fourth determination unit, configured to determine a score corresponding to the feedback information and compare the score with a threshold; a fourth acquisition unit, configured to, when the score is higher than the threshold, acquire target voice information corresponding to the target push message, and add the target push message and the target voice information to the second sample data.

[0074] The above information push device includes a processor and a memory. The above first acquisition unit 21, analysis unit 22, first determination unit 23, sending unit 24, etc. are all stored in the memory as program units, and the processor executes the above program units stored in the memory to implement corresponding functions.

[0075] The processor includes a kernel, and the kernel is used to retrieve corresponding program units from the memory. One or more kernels can be set, and by adjusting the kernel parameters, the problem in the related art that it is difficult to accurately push information for diverse reply messages of different users when performing product push can be solved.

[0076] The memory may include non-permanent memory in the form of computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0077] An embodiment of the present invention provides a computer-readable storage medium with a program stored thereon, and when the program is executed by a processor, the information pushing method is implemented.

[0078] An embodiment of the present invention provides a processor for running a program, and when the program runs, the information pushing method is executed.

[0079] As Figure 3 As shown, an embodiment of the present invention provides an electronic device. The electronic device 30 includes a processor, a memory, and a program stored on the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining call voices generated during the process of pushing products to a user; performing voice analysis on the call voices through a language analysis model to obtain voice information corresponding to the call voices, where the voice information includes semantic information and emotional information; determining push information corresponding to the call voices according to the semantic information and the emotional information to obtain target push information; and sending the target push information to the user. The device herein may be a server, a PC, a PAD, a mobile phone, etc.

[0080] The present application also provides a computer program product, which when executed on a data processing device, is adapted to execute a program initialized with the following method steps: obtaining call voices generated during the process of pushing products to a user; performing voice analysis on the call voices through a language analysis model to obtain voice information corresponding to the call voices, where the voice information includes semantic information and emotional information; determining push information corresponding to the call voices according to the semantic information and the emotional information to obtain target push information; and sending the target push information to the user.

[0081] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0082] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce means for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in one block or multiple blocks.

[0083] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in one block or multiple blocks.

[0084] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 flow or multiple flows and / or blocks Figure 1 or means for implementing the functions specified in one block or multiple blocks.

[0085] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and a memory.

[0086] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.

[0087] A computer-readable medium includes permanent and non-permanent, removable and non-removable media that can store information by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media 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 technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transitory medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer-readable medium does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0088] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or apparatus comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or apparatus. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or apparatus comprising the element.

[0089] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. An information push method, characterized in that, it includes: Obtain the call voice generated during the process of pushing products to users; Perform voice analysis on the call voice through a language analysis model to obtain the voice information corresponding to the call voice, wherein the voice information includes semantic information and emotional information; Determine the push information corresponding to the call voice according to the semantic information and the emotional information to obtain the target push information; Send the target push information to the user; The language analysis model includes a speech recognition sub-model. Performing voice analysis on the call voice through the language analysis model to obtain the voice information corresponding to the call voice includes: Performing semantic recognition on the call voice through the speech recognition sub-model to obtain semantic information; Performing emotional analysis on the semantic information to obtain emotional information; Combining the semantic information and the emotional information to obtain the voice information; The semantic information includes content information and voice information. The language analysis model further includes a dictionary library and an emotional analysis sub-model. Performing emotional analysis on the semantic information to obtain emotional information includes: Determine multiple keywords from the content information, and match the multiple keywords with the sample keywords in the dictionary library to obtain the first emotional information corresponding to each keyword, wherein the dictionary library includes multiple sample keywords and the sample first emotional information corresponding to each sample keyword; Input the multiple keywords and the voice information into the emotional analysis sub-model, and process to obtain the second emotional information, wherein the emotional analysis sub-model is trained by the sample keywords, sample voice information, and sample second emotional information; Combine the first emotional information and the second emotional information to obtain the emotional information.

2. The method according to claim 1, characterized in that, Before inputting the multiple keywords and the voice information into the emotional analysis sub-model and processing to obtain the second emotional information, the method further includes: Obtain multiple sample keywords from the dictionary library, and determine the sample voice information corresponding to each sample keyword and the sample second emotional information corresponding to each sample keyword; Determine each sample keyword, the sample voice information corresponding to the sample keyword, and the sample second emotional information corresponding to the sample keyword as a set of first sample data, and obtain multiple sets of the first sample data; Train a first machine learning model through multiple sets of the first sample data to obtain the emotional analysis sub-model.

3. The method according to claim 1, characterized in that, The semantic information further includes scene information and the relationship information between multiple statements. Determining the push information corresponding to the call voice according to the semantic information and the emotional information to obtain the target push information includes: Input the semantic information and the emotional information into a target model to obtain multiple push information corresponding to the voice information, wherein the target model is trained by sample voice information and sample push information; Determine the target push message from multiple push messages according to the relationship information between the scenario information and the multiple statements.

4. The method according to claim 3, wherein, before inputting the semantic information and the emotion information into a target model to obtain multiple push messages corresponding to the voice information, the method further includes: Determining each historical voice information and the historical push message corresponding to the historical voice information as a set of second sample data, and obtaining multiple sets of the second sample data; Training a second machine learning model with multiple sets of the second sample data to obtain the target model.

5. The method according to claim 4, wherein, after sending the target push message to the user, the method further includes: Obtaining feedback information of the user on the target push message; Determining a score corresponding to the feedback information, and comparing the score with a threshold; When the score is higher than the threshold, obtaining the target voice information corresponding to the target push message, and adding the target push message and the target voice information to the second sample data.

6. An information push device, wherein, comprising: A first acquisition unit for acquiring call voices generated during the process of pushing products to users; An analysis unit for performing voice analysis on the call voices through a language analysis model to obtain voice information corresponding to the call voices, wherein the voice information includes semantic information and emotion information; A first determination unit for determining a push message corresponding to the call voice according to the semantic information and the emotion information to obtain a target push message; A sending unit for sending the target push message to the user; The language analysis model includes a speech recognition sub-model, and the analysis unit includes: a recognition module for performing semantic recognition on the call voices through the speech recognition sub-model to obtain semantic information; an analysis module for performing emotion analysis on the semantic information to obtain emotion information; a combination module for combining the semantic information and the emotion information to obtain the voice information; The semantic information includes content information and voice information, the language analysis model further includes a dictionary library and an emotion analysis sub-model, and the analysis module includes: a determination sub-module for determining multiple keywords from the content information, and matching the multiple keywords with sample keywords in the dictionary library to obtain first emotion information corresponding to each keyword, wherein the dictionary library includes multiple sample keywords and sample first emotion information corresponding to each sample keyword; A processing sub-module for inputting the multiple keywords and the voice information into the emotion analysis sub-model for processing to obtain second emotion information, wherein the emotion analysis sub-model is trained with the sample keywords, sample voice information and sample second emotion information; A combination sub-module for combining the first emotion information and the second emotion information to obtain the emotion information.

7. A computer storage medium, wherein, The computer storage medium is used to store a program, wherein when the program runs, it controls the device where the computer storage medium is located to execute the information push method described in any one of claims 1 to 5.

8. An electronic device, characterized in that it includes one or more processors and a memory, and the memory is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the information push method described in any one of claims 1 to 5.

Images