User portrait construction and updating method of information system based on intelligent IVR (Interactive Voice Response)

Abstract

The invention relates to a user portrait construction and updating method of an information system based on intelligent IVR. The method comprises the following steps: step 1, collecting user voice data through an IVR system; step 2, performing voice analysis on the recorded voice interaction data; 3, associating the historical behavior data of the user with the voice data, and establishing a user behavior model; step 4, constructing a user portrait; and 5, updating the user portrait. According to the method, real-time voice interaction between the user and the IVR system is fully utilized, and the emotion and demand of the user can be captured more accurately, so that a more refined user portrait is constructed.

Description

Technical Field

[0001] The present invention belongs to the technical field of user portrait construction, and relates to a user portrait construction and updating method, in particular to a user portrait construction and updating method for an information system based on an intelligent IVR. Background Art

[0002] With the development of information technology, personalized services and recommendation systems are gaining more and more attention.

[0003] The user portrait construction and update methods in the existing technology mainly rely on static data (such as registration information, historical browsing records, etc.) and single behavior data (such as click records, purchase records), with a long update cycle and unable to reflect the real-time changing emotions and needs of users.

[0004] In addition, existing methods make insufficient use of real-time interaction data and fail to fully consider characteristics such as speech speed, emotion, and tone in user voice interactions. They lack accurate capture of users' immediate emotions and potential needs, resulting in insufficient dimensions, lack of dynamics, and poor timeliness of user portraits, making it difficult to achieve high-precision personalized recommendations and services.

[0005] Therefore, the existing methods have problems such as a single data source, weak dynamic update capabilities, and lack of sentiment analysis. There is an urgent need for an innovative method that can utilize real-time voice interaction data to build and update more comprehensive and dynamic user portraits to meet users' real-time personalized needs and improve service accuracy.

[0006] Therefore, in order to solve the above problems, the present invention proposes a user portrait construction and updating method for an information system based on intelligent IVR.

[0007] After searching, no public documents of the prior art identical or similar to the present invention were found. Summary of the invention

[0008] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a method for constructing and updating user portraits of information systems based on intelligent IVR (Interactive Voice Response), which makes full use of the real-time voice interaction between users and IVR systems, can more accurately capture the user's emotions and needs, and thus construct a more refined user portrait.

[0009] The present invention solves the practical problem by adopting the following technical solutions:

[0010] A method for constructing and updating a user portrait of an information system based on an intelligent IVR comprises the following steps:

[0011] Step 1: Collect user voice data through the IVR system;

[0012] Step 2: Perform voice analysis on the recorded voice interaction data;

[0013] Step 3: Associate the user's historical behavior data with the voice data to establish a user behavior model;

[0014] Step 4: Build a user portrait;

[0015] Step 5: Update user profile.

[0016] Moreover, the specific method of step 1 is:

[0017] Through the IVR system, the voice interaction data between the user and the IVR system is recorded, including: the user's words, speaking speed, and emotions.

[0018] Moreover, the specific method of step 2 is:

[0019] Use speech recognition technology to convert voice interaction data into text, and apply natural language processing technology to perform sentiment analysis and semantic understanding to extract users' emotional tendencies and needs.

[0020] Moreover, the specific method of step 3 is:

[0021] The browsing history and purchase history of the user's historical behavior data are associated with the voice interaction data to establish a user behavior model: first, the user's historical behavior data and voice interaction data are converted into multi-dimensional vector representations; then, the multimodal data fusion method is used for association modeling to generate a unified user feature vector expression; at the same time, a user behavior model is constructed based on these feature vectors, through the formula U=f(H,V), where U is the user behavior model, H is the historical behavior feature vector, V is the voice interaction feature vector, and f is the deep learning fusion function.

[0022] Moreover, the specific method of step 4 is:

[0023] Consider the speech analysis results and user behavior model comprehensively to build a user profile, including the user's interests, preferences, and emotional state.

[0024] Moreover, the specific method of step 5 is:

[0025] In every interaction between the user and the IVR system, the user portrait is updated in real time, the model is continuously optimized, and the accuracy and personalization of the portrait are improved.

[0026] Advantages and beneficial effects of the present invention:

[0027] The present invention proposes a method for constructing and updating user portraits of an information system based on intelligent IVR (Interactive Voice Response). The method utilizes speech recognition technology, natural language processing technology, and data analysis methods to extract and analyze user speech, emotion, voice features, and other information in real time through voice interaction between the user and the IVR system, and construct and update user portraits. When a user is talking to the IVR system, the present invention analyzes the user's voice expression, speech speed, emotional color, etc., and combines the user's historical behavior data to dynamically adjust and optimize the user portrait, thereby realizing personalized services and recommendations.

[0028] 2. Based on intelligent IVR, the present invention realizes real-time voice interaction between users and the system, capturing information such as voice content, emotions, and behavioral data. Through voice recognition and natural language processing, user emotions and needs are analyzed. These are associated with historical behavior data to establish a user behavior model and build a multi-dimensional user portrait, including interests, preferences, emotions, etc. The user portrait is updated in real time for each interaction to improve the level of personalization. The present invention provides enterprises with more accurate and intelligent personalized services, enhancing user satisfaction and market competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a process flow chart of the present invention. DETAILED DESCRIPTION

[0030] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings:

[0031] A method for constructing and updating user portraits of an information system based on intelligent IVR, such as Figure 1 As shown, the following steps are included:

[0032] Step 1: Collect user voice data through the IVR system;

[0033] The specific method of step 1 is:

[0034] Through the IVR system, the voice interaction data between the user and the IVR system is recorded, including: the user's words, speaking speed, emotions, etc.

[0035] Step 2: Perform voice analysis on the recorded voice interaction data;

[0036] The specific method of step 2 is:

[0037] Use speech recognition technology to convert voice interaction data into text, and apply natural language processing technology to perform sentiment analysis and semantic understanding to extract users' emotional tendencies and needs.

[0038] The specific methods of sentiment analysis and semantic understanding include using acoustic signal processing technology to extract speech features (such as intonation, speech speed, volume, MFCCs, etc.), combining deep learning models (such as CNN, LSTM or BERT and acoustic feature fusion model) for analysis, and generating user sentiment scores through sentiment classification (such as positive sentiment score 0.8, negative sentiment score 0.2), thereby identifying the user's emotional tendency (such as positive, negative, neutral) and emotional intensity. At the same time, natural language processing technology is used to perform semantic analysis on the text converted from speech, and through keyword extraction, semantic matching and intent recognition, the user's immediate needs (such as consultation intention, complaint content or purchase intention) are extracted to achieve accurate capture of user emotions and needs.

[0039] Step 3: Associate the user's historical behavior data with the voice data to establish a user behavior model;

[0040] The specific method of step 3 is:

[0041] The user's historical behavior data, such as browsing history, purchase history, and voice interaction data, are associated to establish a user behavior model.

[0042] The association between the user's historical behavior data and voice interaction data is achieved through feature fusion and behavior modeling, specifically using the method of feature vectorization and association analysis. First, the user's historical behavior data (such as browsing history, purchase history) and voice interaction data (such as voice emotion features, semantic intent labels) are converted into multi-dimensional vector representations; then, multimodal data fusion methods (such as Attention mechanism or Transformer model in deep learning) are used for association modeling to generate a unified user feature vector expression. At the same time, a user behavior model is constructed based on these feature vectors, through the formula U = f (H, V), where U is the user behavior model, H is the historical behavior feature vector, V is the voice interaction feature vector, and f is a deep learning fusion function (such as a multi-layer perceptron or RNN).

[0043] Step 4: Build a user portrait;

[0044] The specific method of step 4 is:

[0045] Taking into account the speech analysis results and user behavior model, a user portrait is constructed, including the user's interests, preferences, emotional state, etc.

[0046] The construction of user portraits is based on the specific methods of multi-dimensional data fusion and behavior modeling. First, through feature extraction, the user's voice interaction data (such as speaking speed, sentiment score, semantic intent) and historical behavior data (such as browsing history, purchase history) are converted into structured feature vectors. Next, a deep learning model (such as Transformer or multi-layer perceptron) is used to fuse the features of data from different sources to generate a comprehensive feature expression of the user. Then, cluster analysis and classification algorithms are used to group and label the user's interests, preferences, and emotional states to form a preliminary framework for user portraits. Finally, combined with the correlation analysis of multi-dimensional features, a complete user portrait covering dimensions such as user interests, behavioral habits, and emotional tendencies is generated to support the implementation of personalized services and recommendations.

[0047] Step 5: Update user profile;

[0048] The specific method of step 5 is:

[0049] In every interaction between the user and the IVR system, the user portrait is updated in real time, the model is continuously optimized, and the accuracy and personalization of the portrait are improved.

[0050] Among them, the optimization of user portraits is achieved through real-time updates and dynamic weight adjustments. After each user interacts with the IVR system, the system collects new voice interaction data and behavior data, extracts real-time features and integrates them with existing user portrait features. A dynamic weight allocation method is used to assign different weights according to the timeliness and importance of the data to ensure that new data has a higher influence on the update of the portrait. At the same time, through continuously optimized deep learning models (such as RNN or Transformer), the user behavior model is iteratively updated to capture the changing trends of user needs and emotions, thereby achieving real-time optimization and high-precision dynamic adjustment of user portraits.

[0051] The working principle of the present invention is:

[0052] A method for building and updating user portraits of information systems based on intelligent IVR (Interactive Voice Response) aims to achieve more accurate and personalized user portraits by utilizing real-time voice interaction data, thereby providing personalized services and recommendations. This technology covers the following key aspects:

[0053] Voice data collection: Through the intelligent IVR system, the voice interaction process between the user and the system is recorded. These voice data include the user's words, speaking speed, tone, emotional expression, etc. Voice data collection is the basis for building a refined user portrait.

[0054] Speech analysis: With the help of advanced speech recognition technology, the collected speech data is converted into text form. Then, natural language processing technology is applied to perform sentiment analysis and semantic understanding on the text data to capture the user's emotional tendencies, intentions and needs.

[0055] Historical data integration: Associating the user's historical behavior data (such as browsing history, consultation ticket records, click behavior, etc.) with voice data. This comprehensive analysis helps to establish a user behavior model and provide a more comprehensive information basis for user profiling.

[0056] User portrait construction: Based on the speech analysis results and behavior models, the system will integrate information to build a user portrait. The user portrait includes but is not limited to information about the user's interests, preferences, emotional state, information system usage habits, etc. The user portrait can be regarded as a virtual "label" of the user to help the system better understand the user's needs.

[0057] Profile update and optimization: Every time a user interacts with the IVR system, the system will update the user profile in real time. By continuously accumulating new voice data and behavior data of the user, the profile can be gradually optimized and refined to maintain consistency with user needs.

[0058] The present invention has a wide range of potential applications and can be used in various information systems, such as customer service hotlines, e-commerce platforms, and intelligent assistants. By collecting and analyzing real-time voice interaction data from users, the system can better understand the user's emotions and needs, thereby providing more personalized and intelligent services and recommendations. This technology is expected to promote the business development and competitiveness of enterprises while improving user experience and enhancing user satisfaction.

[0059] It is worth noting that the specific implementation methods and technical details of the present invention can be adjusted and optimized according to actual conditions to achieve better effects and application experience.

[0060] It should be emphasized that the embodiments of the present invention are illustrative rather than restrictive. Therefore, the present invention includes but is not limited to the embodiments described in the specific implementation modes. Any other implementation modes derived by those skilled in the art based on the technical solutions of the present invention also fall within the scope of protection of the present invention.

Claims

1. A method for constructing and updating user portraits of an information system based on intelligent IVR, characterized in that: The following steps are involved: Step 1: Collect user voice data through the IVR system; Step 2: Perform voice analysis on the recorded voice interaction data; Step 3: Associate the user's historical behavior data with the voice data to establish a user behavior model; Step 4: Build a user portrait; Step 5: Update user profile.

2. According to claim 1, a method for constructing and updating user portraits of an information system based on intelligent IVR, characterized in that: The specific method of step 1 is: Through the IVR system, the voice interaction data between the user and the IVR system is recorded, including: the user's words, speaking speed, and emotions.

3. The method for constructing and updating user portraits of an information system based on intelligent IVR according to claim 1, characterized in that: The specific method of step 2 is: Use speech recognition technology to convert voice interaction data into text, and apply natural language processing technology to perform sentiment analysis and semantic understanding to extract users' emotional tendencies and needs.

4. The method for constructing and updating user portraits of an information system based on intelligent IVR according to claim 1, characterized in that: The specific method of step 3 is: The browsing history and purchase history of the user's historical behavior data are associated with the voice interaction data to establish a user behavior model: First, the user's historical behavior data and voice interaction data are converted into multi-dimensional vector representations; then, the multimodal data fusion method is used to perform association modeling to generate a unified user feature vector expression; At the same time, a user behavior model is constructed based on these feature vectors through the formula U=f(H,V), where U is the user behavior model, H is the historical behavior feature vector, V is the voice interaction feature vector, and f is the deep learning fusion function.

5. The method for constructing and updating user portraits of an information system based on intelligent IVR according to claim 1, characterized in that: The specific method of step 4 is: Consider the speech analysis results and user behavior model comprehensively to build a user profile, including the user's interests, preferences, and emotional state.

6. The method for constructing and updating user portraits of an information system based on intelligent IVR according to claim 1, characterized in that: The specific method of step 5 is: In every interaction between the user and the IVR system, the user portrait is updated in real time, the model is continuously optimized, and the accuracy and personalization of the portrait are improved.

Images