A data enhancement method, system and device for directional social anxiety features

Abstract

The present application discloses a data enhancement method, system and device for directional social anxiety features, including: obtaining a first conversation text of a user, the first conversation text at least including text data of a conversation that occurred when the user was in an anxious state; obtaining a second conversation text based on the first conversation text, the second conversation text at least including text data of a conversation that occurred when the user was not in an anxious state and whose semantic similarity with the first conversation text exceeds a first threshold; obtaining at least one anxiety relief feature based on the first conversation text and the second conversation text, the anxiety relief feature being configured as a difference feature between the first conversation text and the second conversation text; generating training content based on the anxiety relief feature. The present invention solves the problem that the training content of the existing solution lacks pertinence, wherein the training content may not be targeted at the actual cause of social anxiety in the user, resulting in poor training effect, and improves the training effect for social anxiety.

Description

Technical Field

[0001] The present application relates to the field of big data technology, and in particular to a data enhancement method, system and device for directional social anxiety characteristics. Background Art

[0002] Social anxiety disorder profoundly affects patients' experience in social situations, causing them to be overly nervous and anxious, and extremely afraid of being noticed, evaluated, or even criticized by others. Patients often pay too much attention to their appearance, words, behavior, and performance, and even trigger a series of physiological reactions, such as blushing, trembling hands, a sudden increase in heart rate, excessive sweating, etc., which seriously hinders smooth communication with others.

[0003] In response to this problem, traditional coping methods mainly rely on clinical diagnosis and psychological counseling. However, these methods face many challenges in actual operation: clinical diagnosis is limited by the shortage of medical resources, high diagnostic difficulty and risk of misdiagnosis; and psychological counseling is often difficult to meet the urgent needs of patients due to the single treatment method and long course of treatment. With the development of technology, intelligent solutions have emerged, opening up new paths for the treatment of social anxiety disorder. For example, the virtual training system and method based on virtual agent interaction demonstrated by the patent with application number CN201810909949.1 provides a variety of interaction modes, situational simulations and virtual dialogue partners with different personalities, allowing users to customize interview training according to their own needs, effectively deal with various anxiety situations, achieve personalized training goals, and significantly reduce anxiety and improve social skills after repeated practice. The patent with application number CN202410113848.9 proposes a virtual reality training system for anxiety disorders, which also provides a progressive and personalized training process, and provides practical coping strategies after training to help patients better cope with social anxiety and improve their social skills and self-confidence; through the virtual scene construction module, the system can create virtual daily life scenes according to the patient's anxiety disorder, providing patients with a safe and controllable training environment; patients wear physiological monitoring equipment in this virtual environment, and the physiological data acquisition module collects and marks physiological data in normal and anxious states. Through analysis, the system can objectively and accurately evaluate the patient's anxiety level; at the same time, the system provides patients with solutions to overcome social anxiety disorders in actual social activities, so that training does not just stay in the virtual environment, but transforms the learned skills into real-life applications.

[0004] However, the causes of social anxiety vary from person to person, and may even vary from scenario to scenario or word to word. Many causes of social anxiety may actually be caused by one or two sentences or words in a conversation. The training content of existing solutions lacks specificity. The training content may not target the actual causes of social anxiety in users, resulting in poor training results. Summary of the invention

[0005] The present invention provides a data enhancement method, system and device for directional social anxiety features, which solves the technical problem that the training content of existing solutions lacks pertinence, wherein the training content may not be targeted at the actual causes of social anxiety in users, resulting in poor training effects.

[0006] A data enhancement method for directional social anxiety features, comprising:

[0007] Obtaining a first conversation text of a user, wherein the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state;

[0008] According to the first dialogue text, a second dialogue text is obtained, wherein the second dialogue text at least includes text data of a dialogue occurring when the user is not in an anxious state and has a semantic similarity with the first dialogue text exceeding a first threshold;

[0009] Obtaining at least one anxiety relief feature according to the first dialogue text and the second dialogue text, wherein the anxiety relief feature is configured as a difference feature between the first dialogue text and the second dialogue text;

[0010] Based on the anxiety relief feature, training content is generated, the training content including at least a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the content of the third dialogue text according to the anxiety relief feature, the semantic similarity between the fourth dialogue text and the third dialogue text exceeding the first threshold.

[0011] Optionally, the obtaining of a first conversation text of the user, wherein the first conversation text at least includes text data of a conversation occurring when the user is in an anxious state, includes:

[0012] In response to detecting that the physiological index of the user meets the preset anxiety index, obtaining the conversation data of the user from before the physiological index changes to after the physiological index recovers as the first conversation data;

[0013] The first conversation data is identified according to the first conversation data to obtain a first conversation text of the user.

[0014] Optionally, obtaining a second dialogue text according to the first dialogue text includes:

[0015] According to the first conversation text, obtaining at least one keyword by keyword extraction;

[0016] According to the at least one keyword, obtaining a conversation including the at least one keyword when the physiological indicator of the user does not meet the preset anxiety indicator as second conversation data;

[0017] According to the second conversation data, the second conversation data is identified and similarity is determined for the first conversation text;

[0018] When the similarity with the first conversation text is greater than a second threshold, a second conversation text of the user is obtained according to the result of recognizing the second conversation data.

[0019] Optionally, obtaining at least one anxiety relief feature according to the first dialogue text and the second dialogue text includes:

[0020] According to the first dialogue text and the second dialogue text, a first word vector group of the first dialogue text and a second word vector group of the second dialogue text are obtained by using a word embedding model;

[0021] According to the first word vector group and the second word vector group, obtaining a vector difference between the first word vector group and the second word vector group;

[0022] At least one anxiety relief feature is obtained according to the vector difference.

[0023] Optionally, obtaining a vector difference between the first word vector group and the second word vector group according to the first word vector group and the second word vector group includes:

[0024] Obtaining grammatical analysis data according to the first dialogue text and the second dialogue text through dependency syntax analysis and / or phrase structure analysis;

[0025] According to the grammatical analysis data, semantic role labeling is performed on the first dialogue text and the second dialogue text to obtain semantic role labeling data;

[0026] Obtaining a corresponding mapping of word vectors of the first word vector group and the second word vector group according to the grammatical analysis data and the semantic role labeling data;

[0027] According to the corresponding mapping between the first word vector group, the second word vector group and the word vector, a vector difference between the first word vector group and the second word vector group is obtained.

[0028] Optionally, generating training content according to the anxiety relief feature, the training content at least including a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the content of the third dialogue text according to the anxiety relief feature, includes:

[0029] Obtaining, based on the first dialogue text, at least one third dialogue text similar to the first dialogue text;

[0030] Obtaining a fourth dialogue text according to the third dialogue text and the anxiety relief feature;

[0031] Training content is generated based on the third dialogue text and the fourth dialogue text.

[0032] Optionally, obtaining a fourth dialogue text according to the third dialogue text and the anxiety relief feature includes:

[0033] According to the third dialogue text, obtaining grammatical analysis data through dependency syntax analysis and / or phrase structure analysis;

[0034] According to the grammatical analysis data, semantic role labeling is performed on the third dialogue text to obtain semantic role labeling data;

[0035] According to a preset semantic role modification order, based on the anxiety relief feature, the third dialogue text is modified until the modified text has a greater similarity to the second dialogue text than to the first dialogue text, and the modified text is used as the fourth dialogue text.

[0036] On the other hand, a data enhancement system for targeted social anxiety features includes a speech recognition platform and a content generation platform;

[0037] The speech recognition platform is configured to:

[0038] Collecting user's conversation information, and identifying the user's conversation information as a first conversation text or a second conversation text and sending it to the content generation platform;

[0039] The content generation platform is configured to:

[0040] Obtaining a first conversation text of a user, wherein the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state;

[0041] According to the first dialogue text, a second dialogue text is obtained, wherein the second dialogue text at least includes text data of a dialogue occurring when the user is not in an anxious state and has a semantic similarity with the first dialogue text exceeding a first threshold;

[0042] Obtaining at least one anxiety relief feature according to the first dialogue text and the second dialogue text, wherein the anxiety relief feature is configured as a difference feature between the first dialogue text and the second dialogue text;

[0043] Based on the anxiety relief feature, training content is generated, the training content including at least a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the content of the third dialogue text according to the anxiety relief feature, the semantic similarity between the fourth dialogue text and the third dialogue text exceeding the first threshold.

[0044] Optionally, it also includes a physiological indicator monitoring platform;

[0045] The physiological index monitoring platform is configured as follows:

[0046] Collect the user's physiological indicators and determine whether the user's physiological indicators meet the preset anxiety indicators;

[0047] When the user's physiological indicators meet the preset anxiety indicators, a trigger instruction is sent to the voice recognition platform;

[0048] The speech recognition platform is also configured to:

[0049] In response to receiving the trigger instruction, collecting the user's conversation information, identifying the user's conversation information as a first conversation text and sending it to the content generation platform;

[0050] In response to receiving the trigger instruction, the user's conversation information is collected, and the user's conversation information is identified as a second conversation text and sent to the content generation platform.

[0051] On the other hand, a computer device is provided. The computer device includes a memory and a processor. The memory stores a computer program. The processor executes the computer program to implement the above method.

[0052] On the other hand, a computer storage medium stores a computer program, and a processor executes the computer program to implement the above method.

[0053] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0054] The present invention discloses a data enhancement method, system and device for directional social anxiety features, comprising: obtaining a first conversation text of a user, wherein the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state; obtaining a second conversation text based on the first conversation text, wherein the second conversation text at least includes text data of a conversation that occurs when the user is not in an anxious state and whose semantic similarity with the first conversation text exceeds a first threshold; obtaining at least one anxiety relief feature based on the first conversation text and the second conversation text, wherein the anxiety relief feature is configured as a difference feature between the first conversation text and the second conversation text; and generating training content based on the anxiety relief feature. At least the problem that the training content of the existing solution lacks pertinence, wherein the training content may not be targeted at the actual cause of social anxiety in the user, resulting in poor training effect, is solved. The present invention combines grammatical structure and deep semantic analysis, and can more accurately capture key information and important intentions in the text. This has important application significance for information extraction, information retrieval and text understanding in natural language processing tasks, and improves the training effect for social anxiety. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the specific implementation of the present application or the technical solution in the prior art, the following is a brief introduction to the drawings required for the specific implementation or the prior art description. In all drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual scale.

[0056] Figure 1 The figure is a flow chart of a data enhancement method for directional social anxiety features of the present invention.

[0057] Figure 2 The figure is a schematic diagram of the structure of a computer device according to the present invention.

[0058] Markings in the figure: 101 - processor, 102 - communication bus, 103 - network interface, 104 - user interface, 105 - memory.

[0059] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0060] In order to enable those skilled in the art to better understand the disclosed solution, the technical solution in the disclosed embodiment will be clearly and completely described below in conjunction with the drawings in the disclosed embodiment. Obviously, the described embodiment is only a part of the disclosed embodiment, not all of the embodiments. According to the embodiments in the disclosed embodiment, all other embodiments obtained by ordinary technicians in the field without creative work should fall within the scope of protection of the disclosed embodiment.

[0061] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.

[0062] Example 1

[0063] like Figure 1 As shown, a data enhancement method for directional social anxiety features includes:

[0064] Step S1, obtaining a first conversation text of a user;

[0065] Specifically, the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state.

[0066] Optionally, the method for obtaining the user's first conversation text includes:

[0067] Based on physiological sensors, such as heart rate monitors, galvanic skin response sensors, and electroencephalograms, the user's physiological indicators are monitored in real time. Generally speaking, the changes in the user's physiological indicators when they are anxious conform to certain rules;

[0068] While the user's physiological indicators are monitored in real time, the user's conversation is recorded, and a first conversation text is obtained through voice recording, text transcription or instant text analysis.

[0069] Optionally, the method for obtaining the user's first conversation text may also include: integrating and analyzing the collected physiological indicator data and conversation data, which can compare the correlation between the changes in physiological indicators and the conversation content through timestamps, so as to obtain the first conversation text without relying on physiological indicator monitoring equipment.

[0070] Step S2, obtaining a second dialogue text according to the first dialogue text;

[0071] Specifically, the second dialogue text at least includes text data of a dialogue that occurs when the user is not in an anxious state and whose semantic similarity with the first dialogue text exceeds a first threshold.

[0072] Optionally, based on the first conversation text, a method for obtaining the second conversation text is to trigger the recording of the user's conversation based on keywords in the first conversation text, and identify the recorded conversation text; when the similarity between the recorded conversation text and the first conversation text exceeds a first threshold, the recorded conversation text is used as the second conversation text.

[0073] Optionally, one user may correspond to one or more first dialogue texts, and one first dialogue text may correspond to one or more second dialogue texts.

[0074] S3, obtaining at least one anxiety relief feature according to the first dialogue text and the second dialogue text;

[0075] Specifically, the anxiety relief feature is configured as a difference feature between the first dialogue text and the second dialogue text.

[0076] Optionally, at least one anxiety relief feature is obtained based on the word vector difference between the first dialogue text and the second dialogue text.

[0077] S4. Generate training content based on anxiety relief characteristics;

[0078] Specifically, the training content includes at least a third dialogue text that makes the user anxious and matching the third dialogue text and a fourth dialogue text feature according to anxiety relief features, and the semantic similarity between the fourth dialogue text and the third dialogue text exceeds a first threshold.

[0079] Optionally, the method for generating training content based on anxiety relief characteristics includes: generating a third dialogue text that simulates a scenario of the first dialogue text based on the first dialogue text, generating training content including the third dialogue text based on the third dialogue text, and simultaneously generating a fourth dialogue text based on the third dialogue text and the anxiety relief characteristics, and establishing an association between the third dialogue text and the fourth dialogue text in the training content based on the fourth dialogue text.

[0080] Through the above scheme, the possibility of users experiencing anxiety when facing conversations with the same substantive content due to differences in wording or other factors can be alleviated, and an association can be established in the user's consciousness between keywords, tones or other words that cause anxiety and keywords, tones or other words that do not cause anxiety, which solves the problem that the training content of existing solutions lacks specificity, and the training content may not be targeted at the actual reasons why users have social anxiety, resulting in poor training results.

[0081] Example 2

[0082] A data enhancement method for directional social anxiety features, comprising:

[0083] S1. Obtain the user's first conversation text;

[0084] Specifically, the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state.

[0085] Optionally, the method for obtaining the user's first conversation text includes:

[0086] Based on physiological sensors, such as heart rate monitors, galvanic skin response sensors, and electroencephalograms, the user's physiological indicators are monitored in real time. Generally speaking, the changes in the user's physiological indicators when they are anxious conform to certain rules;

[0087] While the user's physiological indicators are monitored in real time, the user's conversation is recorded, and a first conversation text is obtained through voice recording, text transcription or instant text analysis.

[0088] Optionally, the method for obtaining the user's first conversation text may also include: integrating and analyzing the collected physiological indicator data and conversation data, which can compare the correlation between the changes in physiological indicators and the conversation content through timestamps, so as to obtain the first conversation text without relying on physiological indicator monitoring equipment.

[0089] Optionally, a first conversation text of the user is obtained, where the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state, including:

[0090] Collect and roll-store user conversation data;

[0091] In response to detecting that the physiological index of the user meets the preset anxiety index, obtaining the conversation data of the user from before the physiological index changes to after the physiological index recovers as the first conversation data;

[0092] According to the first conversation data, the first conversation data is identified to obtain a first conversation text of the user.

[0093] S2. Obtaining a second dialogue text according to the first dialogue text;

[0094] Specifically, the second dialogue text at least includes text data of a dialogue that occurs when the user is not in an anxious state and whose semantic similarity with the first dialogue text exceeds a first threshold.

[0095] Optionally, based on the first conversation text, a method for obtaining the second conversation text is to trigger the recording of the user's conversation based on keywords in the first conversation text, and identify the recorded conversation text; when the similarity between the recorded conversation text and the first conversation text exceeds a first threshold, the recorded conversation text is used as the second conversation text.

[0096] Optionally, one user may correspond to one or more first dialogue texts, and one first dialogue text may correspond to one or more second dialogue texts.

[0097] Optionally, obtaining the second dialogue text according to the first dialogue text includes:

[0098] According to the first conversation text, obtaining at least one keyword by keyword extraction;

[0099] According to the at least one keyword, obtaining a conversation including the at least one keyword when the physiological indicator of the user does not meet the preset anxiety indicator as second conversation data;

[0100] According to the second conversation data, the second conversation data is identified and similarity is determined for the first conversation text;

[0101] When the similarity with the first conversation text is greater than a second threshold, the second conversation text of the user is obtained according to the result of recognizing the second conversation data.

[0102] S3, obtaining at least one anxiety relief feature according to the first dialogue text and the second dialogue text;

[0103] Specifically, the anxiety relief feature is configured as a difference feature between the first dialogue text and the second dialogue text.

[0104] Optionally, at least one anxiety relief feature is obtained based on the word vector difference between the first dialogue text and the second dialogue text.

[0105] Optionally, at least one anxiety relief feature is obtained according to the first dialogue text and the second dialogue text, including:

[0106] According to the first dialogue text and the second dialogue text, a first word vector group of the first dialogue text and a second word vector group of the second dialogue text are obtained through a word embedding model; the word embedding model is Word2Vec, GloVe or FastText.

[0107] According to the first word vector group and the second word vector group, obtaining a vector difference between the first word vector group and the second word vector group;

[0108] At least one anxiety relief feature is obtained based on the vector difference.

[0109] Optionally, obtaining a vector difference between the first word vector group and the second word vector group according to the first word vector group and the second word vector group includes:

[0110] Obtaining grammatical analysis data according to the first dialogue text and the second dialogue text through dependency syntax analysis and / or phrase structure analysis;

[0111] According to the grammatical analysis data, semantic role labeling is performed on the first dialogue text and the second dialogue text to obtain semantic role labeling data;

[0112] Obtaining a corresponding mapping of word vectors of the first word vector group and the second word vector group according to the grammatical analysis data and the semantic role labeling data;

[0113] According to the first word vector group, the second word vector group and the word vector corresponding mapping, a vector difference between the first word vector group and the second word vector group is obtained.

[0114] S4. Generate training content based on anxiety relief characteristics;

[0115] Specifically, the training content includes at least a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the content of the third dialogue text according to anxiety relief characteristics, and the semantic similarity between the fourth dialogue text and the third dialogue text exceeds a first threshold.

[0116] Optionally, the method for generating training content based on anxiety relief characteristics includes: generating a third dialogue text that simulates a scenario of the first dialogue text based on the first dialogue text, generating training content including the third dialogue text based on the third dialogue text, and simultaneously generating a fourth dialogue text based on the third dialogue text and the anxiety relief characteristics, and establishing an association between the third dialogue text and the fourth dialogue text in the training content based on the fourth dialogue text.

[0117] Optionally, training content is generated according to the anxiety relief feature, and the training content includes at least a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the content of the third dialogue text according to the anxiety relief feature, including:

[0118] According to the first dialogue text, obtaining at least one third dialogue text similar to the first dialogue text;

[0119] According to the third dialogue text and the anxiety relief feature, a fourth dialogue text is obtained;

[0120] Training content is generated based on the third dialogue text and the fourth dialogue text.

[0121] Optionally, according to the third dialogue text and the anxiety relief feature, a fourth dialogue text is obtained, including:

[0122] According to the third dialogue text, obtaining grammatical analysis data through dependency syntax analysis and / or phrase structure analysis;

[0123] Performing semantic role labeling on the third dialogue text according to the grammatical analysis data to obtain semantic role labeling data;

[0124] According to the preset semantic role modification order and based on the anxiety relief feature, the third dialogue text is modified until the similarity between the modified text and the second dialogue text is greater than the similarity between the modified text and the first dialogue text, and the modified text is used as the fourth dialogue text.

[0125] Optional methods for determining text similarity include:

[0126] Preprocessing the first text and the second text according to the first text and the second text; the preprocessing includes: word segmentation, stop word removal, case processing, etc.;

[0127] According to the preprocessed first text and the second text, word vectors of the first text and the second text are obtained through a word embedding model; according to word vectorization, a pre-trained word embedding model such as Word2Vec, GloVe or FastText is used to map each word into a word vector;

[0128] The word vectors of all words in the first text and the second text are averaged or weighted averaged to obtain the vector representation of the entire text;

[0129] Perform similarity calculation based on the word vectors of the first text and the second text, such as using cosine similarity or other similarity measurement methods to calculate the similarity between the word vectors of the first text and the second text as the similarity between the first text and the second text;

[0130] The first text may be a first dialogue text, a second dialogue text, a third dialogue text or a fourth dialogue text, and the second text may be a first dialogue text, a second dialogue text, a third dialogue text or a fourth dialogue text that is different from the first text.

[0131] Optionally, semantic role labeling is to further label the semantic roles played by each component in the sentence based on grammatical analysis.

[0132] For the sentence "What is your current job?", semantic role labeling can label:

[0133] You: Agent: This word indicates the performer or subject of the action;

[0134] Present: Time (time role): indicates the time when the action occurs;

[0135] Work: Predicate: indicates the core action or event of a sentence;

[0136] Yes: Copula: a verb that connects the subject and the predicate;

[0137] What: Theme: The specific object of an action or the content of an event; the results of semantic role labeling help to understand the semantics of a sentence more accurately, not just the grammatical structure.

[0138] Combining the results of grammatical analysis and semantic role labeling, the following steps of keyword extraction can be performed:

[0139] Extract core verbs and noun phrases:

[0140] Extract the core verbs and noun phrases in the sentence from the grammatical analysis. These words usually reflect the main action and theme of the sentence.

[0141] Extract the predicate verb and its modifications:

[0142] Identify the most important action verb in the sentence, taking into account its modifiers, such as adjectives and adverbs.

[0143] Semantic role keywords:

[0144] According to the results of semantic role labeling, the components that play important semantic roles in the sentence are identified, such as agent, patient, etc. These components are usually very important in understanding the theme and core intention of the sentence.

[0145] Contextual keywords:

[0146] Consider the modifiers and the subject of the modifier in the sentence, and combine the contextual information of the entire text. For example, for "My current job is a teacher", the keywords may include "I", "now", "work", "teacher", because they directly involve important information such as identity, time, and occupation.

[0147] This method combines grammatical structure and deep semantic analysis to more accurately capture the key information and important intentions in the text. This has important application significance for information extraction, information retrieval and text understanding in natural language processing tasks.

[0148] Optionally, a method for obtaining a preset semantic role modification order includes:

[0149] The entire first conversation text of the user is obtained, and grammatical analysis and semantic role labeling are performed on the entire first conversation text of the user. According to the results of the grammatical analysis and semantic role labeling, a semantic role sequence is obtained. The more times a semantic role appears in the results of the grammatical analysis and semantic role labeling, the higher its ranking in the semantic role sequence. According to the semantic role sequence, the semantic role sequence is used as a preset semantic role modification order.

[0150] Optionally, the semantic role sequence may be screened according to the semantic role sequence, such as removing prepositions, to obtain a preset semantic role modification order.

[0151] Example 3

[0152] A data enhancement system for targeted social anxiety features, including a speech recognition platform and a content generation platform;

[0153] The speech recognition platform is configured as:

[0154] Collecting user conversation information, and identifying the user conversation information as a first conversation text or a second conversation text and sending it to a content generation platform;

[0155] The content generation platform is configured to:

[0156] Obtaining a first conversation text of the user, the first conversation text at least including text data of a conversation that occurs when the user is in an anxious state;

[0157] According to the first dialogue text, a second dialogue text is obtained, wherein the second dialogue text at least includes text data of a dialogue occurring when the user is not in an anxious state and has a semantic similarity with the first dialogue text exceeding a first threshold;

[0158] According to the first dialogue text and the second dialogue text, obtaining at least one anxiety relief feature, wherein the anxiety relief feature is configured as a difference feature between the first dialogue text and the second dialogue text;

[0159] According to the anxiety relief feature, training content is generated, and the training content includes at least a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the content of the third dialogue text according to the anxiety relief feature, and the semantic similarity between the fourth dialogue text and the third dialogue text exceeds a first threshold.

[0160] Optionally, it also includes a physiological indicator monitoring platform;

[0161] The physiological indicator monitoring platform is configured as follows:

[0162] Collect the user's physiological indicators and determine whether the user's physiological indicators meet the preset anxiety indicators;

[0163] When the user's physiological indicators meet the preset anxiety indicators, a trigger command is sent to the voice recognition platform;

[0164] The speech recognition platform is also configured to:

[0165] In response to receiving the trigger instruction, collecting the user's conversation information, identifying the user's conversation information as a first conversation text and sending it to the content generation platform;

[0166] In response to receiving the trigger instruction, the user's conversation information is collected, and the user's conversation information is identified as a second conversation text and sent to the content generation platform.

[0167] Optionally, a first conversation text of the user is obtained, where the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state, including:

[0168] In response to detecting that the physiological index of the user meets the preset anxiety index, obtaining the conversation data of the user from before the physiological index changes to after the physiological index recovers as the first conversation data;

[0169] According to the first conversation data, the first conversation data is identified to obtain a first conversation text of the user.

[0170] Optionally, obtaining the second dialogue text according to the first dialogue text includes:

[0171] According to the first conversation text, obtaining at least one keyword by keyword extraction;

[0172] According to the at least one keyword, obtaining a conversation including the at least one keyword when the physiological indicator of the user does not meet the preset anxiety indicator as second conversation data;

[0173] According to the second conversation data, the second conversation data is identified and similarity is determined for the first conversation text;

[0174] When the similarity with the first conversation text is greater than a second threshold, the second conversation text of the user is obtained according to the result of recognizing the second conversation data.

[0175] Optionally, the second threshold may be the same as the first threshold or may be different from the first threshold, and may be set as required.

[0176] Optionally, at least one anxiety relief feature is obtained according to the first dialogue text and the second dialogue text, including:

[0177] According to the first dialogue text and the second dialogue text, a first word vector group of the first dialogue text and a second word vector group of the second dialogue text are obtained by using a word embedding model;

[0178] According to the first word vector group and the second word vector group, obtaining a vector difference between the first word vector group and the second word vector group;

[0179] At least one anxiety relief feature is obtained based on the vector difference.

[0180] Optionally, obtaining a vector difference between the first word vector group and the second word vector group according to the first word vector group and the second word vector group includes:

[0181] Obtaining grammatical analysis data according to the first dialogue text and the second dialogue text through dependency syntax analysis and / or phrase structure analysis;

[0182] According to the grammatical analysis data, semantic role labeling is performed on the first dialogue text and the second dialogue text to obtain semantic role labeling data;

[0183] Obtaining a corresponding mapping of word vectors of the first word vector group and the second word vector group according to the grammatical analysis data and the semantic role labeling data;

[0184] According to the first word vector group, the second word vector group and the word vector corresponding mapping, a vector difference between the first word vector group and the second word vector group is obtained.

[0185] Optionally, training content is generated according to the anxiety relief feature, and the training content includes at least a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the content of the third dialogue text according to the anxiety relief feature, including:

[0186] According to the first dialogue text, obtaining at least one third dialogue text similar to the first dialogue text;

[0187] According to the third dialogue text and the anxiety relief feature, a fourth dialogue text is obtained;

[0188] Training content is generated based on the third dialogue text and the fourth dialogue text.

[0189] Optionally, according to the third dialogue text and the anxiety relief feature, a fourth dialogue text is obtained, including:

[0190] According to the third dialogue text, obtaining grammatical analysis data through dependency syntax analysis and / or phrase structure analysis;

[0191] Performing semantic role labeling on the third dialogue text according to the grammatical analysis data to obtain semantic role labeling data;

[0192] According to the preset semantic role modification order and based on the anxiety relief feature, the third dialogue text is modified until the similarity between the modified text and the second dialogue text is greater than the similarity between the modified text and the first dialogue text, and the modified text is used as the fourth dialogue text.

[0193] Example 4

[0194] This embodiment provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement any of the above methods.

[0195] Specifically, Figure 2 As shown, Figure 2 This is a structural diagram of a computer device in the present application, and the computer device may include: a processor 101, such as a central processing unit (CPU), a communication bus 102, a user interface 104, a network interface 103, and a memory 105. Among them, the communication bus 102 is used to realize the connection and communication between these components. The user interface 104 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the user interface 104 may also include a standard wired interface and a wireless interface. The network interface 103 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (WIreless-FIdelity, WI-FI) interface). The memory 105 may optionally be a storage device independent of the aforementioned processor 101. The memory 105 may be a high-speed random access memory (RAM) memory, or a stable non-volatile memory (NVM), such as at least one disk storage. The processor 101 may be a general-purpose processor, including a central processing unit, a network processor, etc., or may be a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, or a discrete hardware component.

[0196] Those skilled in the art will appreciate that Figure 2 The structure shown in the figure does not constitute a limitation on the electronic device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0197] like Figure 2 As shown, the memory 105 as a storage medium may include an operating system, a network communication module, a user interface module, and an application program for implementing a data enhancement method for directional social anxiety characteristics.

[0198] exist Figure 2In the electronic device shown, the network interface 103 is mainly used for data communication with a network server; the user interface 104 is mainly used for data interaction with a user; the processor 101 and the memory 105 in the present application can be set in the electronic device, and the electronic device calls the application program for implementing a data enhancement method for directional social anxiety characteristics stored in the memory 105 through the processor 101 to implement the above method.

[0199] Example 5

[0200] This embodiment provides a computer-readable storage medium, on which a computer program is stored, and a processor executes the computer program to implement any of the above methods.

[0201] In some embodiments, the computer readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or various devices including one or any combination of the above memories. The computer may be various computing devices including intelligent terminals and servers.

[0202] In the above embodiments of the present disclosure, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0203] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0204] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0205] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0206] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable non-volatile storage medium. According to this understanding, the technical solution of the present disclosure, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a non-volatile storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the various embodiments of the present disclosure. The aforementioned non-volatile storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc. Various media that can store program codes.

[0207] The above are only preferred embodiments of the present disclosure. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present disclosure. These improvements and modifications should also be regarded as the protection scope of the present disclosure.

Claims

1. A data enhancement method for directional social anxiety features, characterized in that: The steps include: Obtaining a first conversation text of a user, wherein the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state; According to the first dialogue text, a second dialogue text is obtained, wherein the second dialogue text at least includes text data of a dialogue occurring when the user is not in an anxious state and has a semantic similarity with the first dialogue text exceeding a first threshold; Obtaining at least one anxiety relief feature according to the first dialogue text and the second dialogue text, wherein the anxiety relief feature is configured as a difference feature between the first dialogue text and the second dialogue text; Generate training content according to the anxiety relief feature, the training content at least including a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the third dialogue text according to the anxiety relief feature, wherein the semantic similarity between the fourth dialogue text and the third dialogue text exceeds the first threshold; The step of obtaining a second dialogue text according to the first dialogue text includes: According to the first conversation text, obtaining at least one keyword by keyword extraction; According to the at least one keyword, obtaining a conversation including the at least one keyword when the physiological indicator of the user does not meet the preset anxiety indicator as second conversation data; According to the second conversation data, the second conversation data is identified and similarity is determined for the first conversation text; When the similarity with the first conversation text is greater than a second threshold, obtaining a second conversation text of the user according to the result of recognizing the second conversation data; The generating of training content according to the anxiety relief feature, wherein the training content at least includes a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the content of the third dialogue text according to the anxiety relief feature, comprises: Obtaining, based on the first dialogue text, at least one third dialogue text similar to the first dialogue text; Obtaining a fourth dialogue text according to the third dialogue text and the anxiety relief feature; Generate training content according to the third dialogue text and the fourth dialogue text; The step of obtaining a fourth dialogue text according to the third dialogue text and the anxiety relief feature comprises: According to the third dialogue text, obtaining grammatical analysis data through dependency syntax analysis and / or phrase structure analysis; According to the grammatical analysis data, semantic role labeling is performed on the third dialogue text to obtain semantic role labeling data; According to a preset semantic role modification order, based on the anxiety relief feature, the third dialogue text is modified until the modified text has a greater similarity to the second dialogue text than to the first dialogue text, and the modified text is used as the fourth dialogue text.

2. The data enhancement method for directional social anxiety features according to claim 1, characterized in that: The obtaining of the first conversation text of the user, wherein the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state, includes: In response to detecting that the physiological index of the user meets the preset anxiety index, obtaining the conversation data of the user from before the physiological index changes to after the physiological index recovers as the first conversation data; The first conversation data is identified according to the first conversation data to obtain a first conversation text of the user.

3. The data enhancement method for directional social anxiety features according to claim 1, characterized in that: The step of obtaining at least one anxiety relief feature according to the first dialogue text and the second dialogue text includes: According to the first dialogue text and the second dialogue text, a first word vector group of the first dialogue text and a second word vector group of the second dialogue text are obtained by using a word embedding model; According to the first word vector group and the second word vector group, obtaining a vector difference between the first word vector group and the second word vector group; At least one anxiety relief feature is obtained according to the vector difference.

4. The data enhancement method for directional social anxiety features according to claim 3, characterized in that: The obtaining, according to the first word vector group and the second word vector group, a vector difference between the first word vector group and the second word vector group includes: Obtaining grammatical analysis data according to the first dialogue text and the second dialogue text through dependency syntax analysis and / or phrase structure analysis; According to the grammatical analysis data, semantic role labeling is performed on the first dialogue text and the second dialogue text to obtain semantic role labeling data; Obtaining a corresponding mapping of word vectors of the first word vector group and the second word vector group according to the grammatical analysis data and the semantic role labeling data; According to the corresponding mapping between the first word vector group, the second word vector group and the word vector, a vector difference between the first word vector group and the second word vector group is obtained.

5. A data enhancement system for directional social anxiety features, characterized in that: Includes speech recognition platform and content generation platform; The speech recognition platform is configured to: Collecting user's conversation information, and identifying the user's conversation information as a first conversation text or a second conversation text and sending it to the content generation platform; The content generation platform is configured to: Obtaining a first conversation text of a user, wherein the first conversation text at least includes text data of a conversation that occurs when the user is in an anxious state; According to the first dialogue text, a second dialogue text is obtained, wherein the second dialogue text at least includes text data of a dialogue occurring when the user is not in an anxious state and has a semantic similarity with the first dialogue text exceeding a first threshold; Obtaining at least one anxiety relief feature according to the first dialogue text and the second dialogue text, wherein the anxiety relief feature is configured as a difference feature between the first dialogue text and the second dialogue text; Generate training content according to the anxiety relief feature, the training content at least including a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the third dialogue text according to the anxiety relief feature, wherein the semantic similarity between the fourth dialogue text and the third dialogue text exceeds the first threshold; The step of obtaining a second dialogue text according to the first dialogue text includes: According to the first conversation text, obtaining at least one keyword by keyword extraction; According to the at least one keyword, obtaining a conversation including the at least one keyword when the physiological indicator of the user does not meet the preset anxiety indicator as second conversation data; According to the second conversation data, the second conversation data is identified and similarity is determined for the first conversation text; When the similarity with the first conversation text is greater than a second threshold, obtaining a second conversation text of the user according to the result of recognizing the second conversation data; The generating of training content according to the anxiety relief feature, wherein the training content at least includes a third dialogue text that makes the user anxious and a fourth dialogue text obtained by processing the content of the third dialogue text according to the anxiety relief feature, comprises: Obtaining, based on the first dialogue text, at least one third dialogue text similar to the first dialogue text; Obtaining a fourth dialogue text according to the third dialogue text and the anxiety relief feature; Generate training content according to the third dialogue text and the fourth dialogue text; The step of obtaining a fourth dialogue text according to the third dialogue text and the anxiety relief feature comprises: According to the third dialogue text, obtaining grammatical analysis data through dependency syntax analysis and / or phrase structure analysis; According to the grammatical analysis data, semantic role labeling is performed on the third dialogue text to obtain semantic role labeling data; According to a preset semantic role modification order, based on the anxiety relief feature, the third dialogue text is modified until the modified text has a greater similarity to the second dialogue text than to the first dialogue text, and the modified text is used as the fourth dialogue text.

6. The data enhancement system for directional social anxiety features according to claim 5, characterized in that: It also includes a physiological indicator monitoring platform; The physiological index monitoring platform is configured as follows: Collect the user's physiological indicators and determine whether the user's physiological indicators meet the preset anxiety indicators; When the user's physiological indicators meet the preset anxiety indicators, a trigger instruction is sent to the voice recognition platform; The speech recognition platform is also configured to: In response to receiving the trigger instruction, collecting the user's conversation information, identifying the user's conversation information as a first conversation text and sending it to the content generation platform; In response to receiving the trigger instruction, the user's conversation information is collected, and the user's conversation information is identified as a second conversation text and sent to the content generation platform.

7. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and the processor executes the computer program to implement the method according to any one of claims 1 to 4.

Images