Group clue discovery method based on multi-feature fusion

Abstract

The invention provides a group clue discovery method based on multi-feature fusion. Comprising two parts: a key group identification module based on multi-feature fusion and a key group clue discovery module; the key group recognition module based on multi-feature fusion inputs a word embedding sequence subjected to position coding into a multi-layer Transfor mer network, and deep feature extraction is carried out by utilizing a self-attention mechanism; and then designing a multi-expert hybrid model to perform feature fusion and discrimination on code value features, representation of text features and dimensionality-reduced statistical features, and specifically, designing a gating mechanism to analyze input features to generate weighting coefficients of each expert network; and the key group clue discovery module receives an operation result of the multi-expert hybrid model, realizes screening of key groups through a large model prompt-based text clue discovery and sliding window time sequence clue discovery method, and obtains a group screening result.

Description

Technical Field

[0001] The present disclosure relates to the field of computers, and more specifically, to a group clue discovery method based on multi-feature fusion. Background Art

[0002] Due to the complexity, diversity, and real-time characteristics of platform data, the media of streaming media information dissemination in the platform are diversified, and the key information data is sparsely distributed, has a very low proportion, and is discrete and hidden. This increases the difficulty of group identification and obtaining group clues in the platform. In addition, the clue group data in the platform usually appears in the form of multi-dimensional features, such as text, code values, statistical information, etc., and most of the existing methods interactively represent various types of features and extract key clues from multi-dimensional feature data based on a unified representation. However, there are semantic differences between feature data of different categories. Therefore, it is difficult for existing methods to learn multi-dimensional features with strong correlation attributes, and it is impossible to use the intrinsic connection and dependency between features of different categories to assist in the discovery of key clues, which leads to the deterioration of the effect of the model in extracting multi-dimensional features of groups. Summary of the invention

[0003] The purpose of the embodiments of the present disclosure is to provide a group clue discovery method based on multi-feature fusion. The main research goal of the present invention is to identify and analyze key groups more efficiently.

[0004] First, the present invention identifies key groups by comprehensively analyzing multiple features such as text, code values, and statistical information. These features include but are not limited to text semantic features, numerical code features, and statistical features extracted from historical data. Through a multi-dimensional information analysis method, the limitations of relying solely on a single feature are avoided, thereby significantly improving the accuracy and reliability of key group identification.

[0005] After identifying the key groups, the present invention further combines the big model and time series information to conduct in-depth mining of the data in the group to discover hidden important clues. The application of the big model enables the present invention to process and understand complex patterns and relationships, while the analysis of time series information can capture the dynamic changes of data in the time dimension, thereby identifying clues that stand out at specific time points or time periods. In this way, the present invention can not only discover potential important information, but also predict and explain the evolution trend of this information, providing strong support for subsequent decision-making.

[0006] In a general aspect, a group clue discovery method based on multi-feature fusion is provided, comprising two parts: a key group identification module based on multi-feature fusion and a key group clue discovery module;

[0007] The key group identification module based on multi-feature fusion first uses the GPT model to extract group text data on the social platform, and pre-processes the text data to convert the text data into word embedding representation;

[0008] Next, the word embedding sequence after position encoding is input into the multi-layer Transformer network to perform deep feature extraction using the self-attention mechanism. The calculation formula is:

[0009]

[0010] Where Q represents the query matrix, which indicates the degree of attention of the current word to other words. A query vector is generated for each word position to find words related to it. K represents the key matrix, which represents the characteristics or content of the word. A key vector is generated for each word position to match with the query vector (Q) to determine the relevance. V represents the value matrix, which represents the value or output information of the word. A value vector is generated for each word position, which represents the content information of the word, and will be weighted and summarized into the output. d k Represents the dimension of the key vector K, which is a normalization factor used to reduce the value range of the dot product result to avoid excessively large values ​​causing the softmax function to approach saturation and thus lose discrimination. T represents the transposed matrix of the key matrix K; the deep features include code value features, text features, and statistical features;

[0011] The multi-expert hybrid model first establishes independent expert networks for text features, code value features and statistical features, each of which is a multi-layer perceptron, which generates representations for text features and reduces the dimension of statistical features; then the multi-expert hybrid model is designed to perform feature fusion and discrimination on the code value features, the representation of text features, and the statistical features after dimensionality reduction. Specifically, a gating mechanism is designed to generate weighted coefficients for each expert network by analyzing the input features, and dynamically adjust the output weights of each expert network according to the importance of the input features; finally, the outputs of all expert networks are fused by weighted summation to generate a comprehensive feature representation, and the fusion formula is specifically as follows:

[0012]

[0013] Where n is the number of expert networks, F final is the final feature representation after fusion;

[0014] The key group clue discovery module receives the operation results of the multi-expert hybrid model, and screens key groups through text clue discovery based on large model prompts and sliding window time series clue discovery methods, and obtains group screening results.

[0015] The analysis method of the code value feature is: focusing on code values ​​that appear repeatedly in multiple groups or have obvious directionality;

[0016] The statistical feature analysis method is: using a multi-layer perceptron model to fuse all statistical features to generate the final statistical feature representation. The calculation formula is as follows:

[0017] F stat =σ(W×X+b)

[0018] Among them, X is the statistical feature vector of the input, W is the weight matrix, b is the bias, and σ is the activation function.

[0019] The text clue discovery based on large model prompts utilizes a pre-trained language model and designs a suitable prompt. When processing the prompt, the model comprehensively considers the word frequency, semantic association and emotional expression in the text to generate an output containing key information, thereby mining text clues closely related to group behavior, including the group's charter, emotional tendencies, common topics, hot keywords and communication style.

[0020] The sliding window time series clue discovery method dynamically analyzes the time dimension of group activities on social platforms to identify specific behaviors. Specifically, the data is divided into overlapping time periods, and the group's speech frequency, keyword usage, and user activity indicators are independently analyzed in each time period to capture sudden behaviors and hot activities within the group.

[0021] The similarity of the activity patterns of two groups in different time periods is judged. The dynamic time rule method is introduced to measure the time series similarity of the two groups. By calculating the minimum distance between the activity patterns of the groups, the groups with similar activity characteristics in different time periods are identified. The calculation formula is:

[0022]

[0023] Among them, X[i] and Y[j] represent the values ​​of two time series X and Y at the i-th and j-th time points respectively, and n is the length of the time series

[0024] The innovation of the embodiment of the present invention lies in:

[0025] The present invention improves the accuracy and response speed of group identification in multiple dimensions. By combining text, code value and statistical features, a comprehensive analysis of the group is achieved, effectively solving the limitations of single feature analysis in the prior art. In addition, the time series analysis module can capture group activities in hot events in a timely manner, with higher hotness and real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The above and other objects and features of the present disclosure will become more apparent from the following description in conjunction with the accompanying drawings.

[0027] Figure 1 is a schematic diagram showing a multi-expert mixture model according to an embodiment of the present disclosure;

[0028] Figure 2 is a schematic diagram showing an architecture diagram of an open source code repository summary automatic generation model according to an embodiment of the present disclosure;

[0029] Figure 3 is a schematic diagram showing dynamic time warping according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0030] The following specific embodiments are provided to help the reader obtain a comprehensive understanding of the methods, devices and / or systems described herein. However, after understanding the disclosure of the present application, various changes, modifications and equivalents of the methods, devices and / or systems described herein will be clear. For example, the order of operations described herein is only an example and is not limited to those orders set forth herein, but can be changed as will be clear after understanding the disclosure of the present application, except for operations that must occur in a specific order. In addition, for greater clarity and simplicity, the description of features known in the art may be omitted.

[0031] The features described herein can be implemented in different forms and should not be construed as being limited to the examples described herein. Rather, the examples described herein have been provided to illustrate only some of the many possible ways to implement the methods, devices, and / or systems described herein, which will be clear after understanding the disclosure of the present application.

[0032] As used herein, the term "and / or" includes any one of the associated listed items and any combination of any two or more.

[0033] Although terms such as "first", "second", and "third" may be used herein to describe various members, components, regions, layers, or portions, these members, components, regions, layers, or portions should not be limited by these terms. Instead, these terms are only used to distinguish one member, component, region, layer, or portion from another member, component, region, layer, or portion. Therefore, without departing from the teachings of the examples described herein, the first member, first component, first region, first layer, or first portion referred to in the examples may also be referred to as the second member, second component, second region, second layer, or second portion.

[0034] In the specification, when an element (such as a layer, a region or a substrate) is described as being “on”, “connected to” or “coupled to” another element, the element may be directly “on”, “connected to” or “coupled to” another element, or one or more other elements may be present therebetween. Conversely, when an element is described as being “directly on”, “directly connected to” or “directly coupled to” another element, there may be no other elements present therebetween.

[0035] The terms used herein are only used to describe various examples and are not intended to limit the disclosure. Unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. The terms "comprise", "include" and "have" indicate the presence of the described features, quantities, operations, components, elements and / or combinations thereof, but do not exclude the presence or addition of one or more other features, quantities, operations, components, elements and / or combinations thereof.

[0036] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by a person of ordinary skill in the art to which the present disclosure belongs after understanding the present disclosure. Unless explicitly defined as such herein, terms (such as those defined in a general dictionary) should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and the present disclosure, and should not be interpreted in an idealized or overly formal manner.

[0037] Furthermore, in the description of examples, when it is considered that a detailed description of a well-known related structure or function would cause vague interpretation of the present disclosure, such a detailed description will be omitted.

[0038] Figure 1 It is a schematic diagram showing a group clue discovery method based on multi-feature fusion according to an embodiment of the present disclosure.

[0039] First, the present invention identifies key groups by comprehensively analyzing multiple features such as text, code values, and statistical information. These features include but are not limited to text semantic features, numerical code features, and statistical features extracted from historical data. Through a multi-dimensional information analysis method, the limitations of relying solely on a single feature are avoided, thereby significantly improving the accuracy and reliability of key group identification.

[0040] After identifying the key groups, the present invention further combines the big model and time series information to conduct in-depth mining of the data in the group to discover hidden important clues. The application of the big model enables the present invention to process and understand complex patterns and relationships, while the analysis of time series information can capture the dynamic changes of data in the time dimension, thereby identifying clues that stand out at specific time points or time periods. In this way, the present invention can not only discover potential important information, but also predict and explain the evolution trend of this information, providing strong support for subsequent decision-making.

[0041] Key group identification based on multi-feature fusion

[0042] In the present invention, the GPT model is used to extract features from group text information on social platforms, including group names, descriptions, chat content, etc. The extraction of text features is an important part of group analysis, which can help reveal the communication patterns and potential interactive behaviors between group members. In order to ensure the consistency of the data and the accuracy of the model, the text data is first preprocessed, including removing irrelevant characters, standardizing punctuation, and processing spelling errors. This step ensures that the data format input into the model is unified and reduces the impact of noise on model performance.

[0043] The preprocessed text data is converted into word embedding representation. Then, the position-encoded word embedding sequence is input into the multi-layer Transformer network for deep feature extraction. The core mechanism of the Transformer network is the self-attention mechanism, which can dynamically pay attention to other parts of the text when processing each word, thereby capturing long-range word dependencies. The calculation formula of the self-attention mechanism is:

[0044]

[0045] In this formula, Q represents the query matrix, K represents the key matrix, V represents the value matrix, and d k Represents the dimension of the key vector. The self-attention mechanism can effectively integrate the correlation between words to generate feature representations rich in contextual information.

[0046] In order to extract and analyze code value features, it is first necessary to identify all possible code value information from the text data of the group. This information may include URL links, hash values, encoded strings, etc. Through rule matching and pattern recognition technology, these key code values ​​can be effectively extracted. The extracted code value information is then further parsed to identify the potential relationships between them. In the process of parsing code value features, special attention is paid to those code values ​​that appear repeatedly in multiple groups or have obvious directionality. This type of information is often the key node of interaction between groups. By analyzing these nodes, the structure of the group network can be more accurately described and clues of hot activities can be identified.

[0047] Statistical features are an important aspect of identifying group behavior patterns. Statistical features cover user behavior data within the group, such as the number of users, speaking frequency, message length, IP distribution, etc. These features can reflect the activity of the group, user composition and its interaction pattern, and help identify hot activities. In order to integrate these statistical information, the present invention uses a multi-layer perceptron (MLP) model to fuse all statistical features and generate the final statistical feature representation. Its core calculation formula is as follows:

[0048] F stat =σ(W×X+b)

[0049] Among them, X is the input statistical feature vector, W is the weight matrix, b is the bias, and σ is the activation function. Through this formula, multiple statistical features can be fused into a high-dimensional representation for subsequent group analysis.

[0050] In order to make full use of the multi-dimensional information of text features, code value features and statistical features, the present invention designs a multi-expert hybrid model for feature fusion and discrimination. This model can flexibly select and combine the most suitable expert network according to the characteristics of different features to improve the recognition accuracy of key groups.

[0051] The core idea of ​​the multi-expert hybrid model is to assign features to different expert networks, each of which focuses on processing a specific type of feature, such as text features, code value features, or statistical features. Then, the outputs of each expert network are weighted and fused through a gating mechanism to obtain a comprehensive feature representation. This approach can effectively deal with the heterogeneity of multi-dimensional feature data and fully explore the potential information of various features, such as Figure 2 as shown.

[0052] First, the text features, code value features, and statistical features are processed by independent expert networks. Each expert network is a multi-layer perceptron (MLP). In this way, each expert network can independently extract a deep representation of the input features.

[0053] Next, the gating mechanism dynamically adjusts the output weights of each expert network based on the importance of the input features. The gating network generates weighted coefficients for each expert network by analyzing the input features, which determine the contribution of each expert network to the final output.

[0054] Finally, the outputs of all expert networks are fused through weighted summation to generate a comprehensive feature representation. The fused feature representation can express the multi-dimensional characteristics of the group, has stronger discriminative power, and can help the model identify key groups more accurately. The fusion formula is as follows:

[0055]

[0056] Where n is the number of expert networks, F final It is the final feature representation after fusion, and the expert model is composed of multi-layer perceptrons stacked together.

[0057] In addition, the multi-expert hybrid model has strong scalability. It can increase or decrease the number of expert networks according to actual application requirements and optimize the structure of each expert network for specific applications. This makes the present invention widely applicable to different types of social platform data analysis, especially in the face of complex and diverse group information, showing strong robustness and adaptability.

[0058] Key group clues found

[0059] Discovering text clues based on large model prompts is one of the key steps in group analysis. By using pre-trained language models (such as GPT) and designing appropriate prompts, text clues closely related to group behavior can be effectively mined. These clues include the group's charter, emotional tendencies, common topics, hot keywords, and communication styles, helping to identify potential key groups.

[0060] In the application of large language models, the design of prompts directly affects the performance of the model. Prompts are the key to guiding the language model to generate target outputs. They need to be carefully designed according to the specific requirements of the task to ensure that the model can accurately capture key information. In this paper, a series of highly targeted prompts are designed to extract important text clues in the group.

[0061] Here is a prompt design for identifying the core topic of a group:

[0062] "Please analyze the following text and list the three most frequently discussed topics, as well as the possible emotional tendencies behind these topics:\n\n'{Group chat history}'\n\nPlease describe each topic and its related keywords in detail."

[0063] This prompt guides the model to not only identify the main topics when analyzing the text, but also to infer the emotional tendencies of these topics based on the context. In this way, the overall atmosphere of the group and its potential risks can be better understood. When processing this prompt, the model will comprehensively consider the vocabulary frequency, semantic associations, and emotional expressions in the text to generate an output containing key information. For example, in a group discussing a hot event, the model may identify the emotional state related to the event and mark some sensitive keywords, which provides an important basis for the subsequent identification of key groups.

[0064] In addition, in order to more comprehensively capture the communication style of the group, Prompt is designed to analyze the characteristics of the conversation:

[0065] "Please outline common expressions in the following conversations, including whether there is bias or slant toward specific events:\n\n'{Group chat log}'\n\nAnd indicate whether there are recurring slogans or specific language patterns."

[0066] Through this prompt, the model can identify language patterns, recurring slogans, and potential biases in conversations. Such analysis can reveal whether the communication style of a group has specific tendencies or other behavioral characteristics, helping to identify potential focus groups in advance.

[0067] The sliding window time series clue discovery method identifies specific behaviors by dynamically analyzing the time dimension of group activities on social platforms. The sliding window technology divides the data into overlapping time periods, and independently analyzes key indicators such as the group's speech frequency, keyword usage, and user activity in each time period, so as to capture sudden behaviors and hot activities within the group in a timely manner. By setting the time granularity of the sliding window to 1 hour or 30 minutes, a rapid response to hot events can be ensured. In each time window, the changes in the current indicator and the previous window can be compared to effectively identify the hot time period. For example, a sudden surge in the number of speeches in a certain time window may indicate that the group is focusing on discussing hot events or planning hot activities.

[0068] Next, we introduce some advanced time series analysis and pattern recognition methods to identify the similarity of the activity patterns of the two groups in different time periods. In particular, the Dynamic Time Warping (DTW) method is used to measure the similarity of the time series of the two groups. DTW is a powerful time series analysis tool that can align and compare time series of different lengths to identify the similarities between them. In the present invention, DTW can effectively identify groups with similar activity characteristics in different time periods by calculating the minimum distance between the activity patterns of the groups. Figure 3shown.

[0069] The calculation formula of DTW is as follows:

[0070]

[0071] Where X[i] and Y[j] represent the values ​​of the two time series X and Y at the i-th and j-th time points, respectively, and n is the length of the time series. DTW aligns the time series and calculates their similarity by calculating the Euclidean distance of all possible pairs between the two time series and finding the pairing path that minimizes the total distance.

[0072] Combining kurtosis analysis and multi-feature fusion strategy, the present invention can accurately locate high-risk time periods and reveal the temporal patterns of group behavior, providing strong support for the comprehensive analysis of key groups.

[0073] In a more specific embodiment, the present invention provides:

[0074] Key group identification module based on multi-feature fusion

[0075] The present invention uses the GPT model to extract features from group text information on social platforms, including group names, descriptions, chat content, etc. The extraction of text features is an important part of group analysis, which can help reveal the communication patterns and potential specific behaviors between group members. In order to ensure the consistency of the data and the accuracy of the model, the text data is first preprocessed, including removing irrelevant characters, standardizing punctuation, and processing spelling errors. This step ensures that the data format input into the model is unified and reduces the impact of noise on model performance.

[0076] The preprocessed text data is converted into word embedding representation. Then, the position-encoded word embedding sequence is input into the multi-layer Transformer network for deep feature extraction. The core mechanism of the Transformer network is the self-attention mechanism, which can dynamically pay attention to other parts of the text when processing each word, thereby capturing long-range word dependencies. The calculation formula of the self-attention mechanism is:

[0077]

[0078] In this formula, Q represents the query matrix, K represents the key matrix, V represents the value matrix, and \(d_k\) represents the dimension of the key vector. The self-attention mechanism can effectively integrate the correlation between words to generate feature representations rich in contextual information.

[0079] If the vector dimension is specified as 512, it will be converted into the following after GPT text preprocessing:

[0080] w1=[0.02133,0.10713,.............0.00159]

[0081] In the group analysis of social platforms, in order to extract and analyze code value features, it is first necessary to identify all possible code value information from the text data of the group. This information may include URL links, hash values, encoded strings, etc. Through rule matching and pattern recognition technology, these key code values ​​can be effectively extracted. The extracted code value information is then further parsed to identify the potential relationship between them. In the process of parsing code value features, special attention is paid to those code values ​​that appear repeatedly in multiple groups or have obvious directionality. This type of information is often the key node of interaction between groups. By analyzing these nodes, the structure of the group network can be more accurately described and clues of hot activities can be identified. Through the analysis of the 'China XX Network' group, it was found that there was a link jumping to "https: / / weibo.com / u / 2377587254"

[0082] Statistical features are an important aspect of identifying group behavior patterns. Statistical features cover user behavior data within the group, such as the number of users, speaking frequency, message length, IP distribution, etc. These features can reflect the activity of the group, user composition and its interaction pattern, and help identify hot activities. In order to integrate these statistical information, the present invention uses a multi-layer perceptron (MLP) model to fuse all statistical features and generate the final statistical feature representation.

[0083] F stat =σ(W×X+b)

[0084] Among them, X is the input statistical feature vector, W is the weight matrix, b is the bias, and σ is the activation function. Through this formula, multiple statistical features can be fused into a high-dimensional representation for subsequent group analysis.

[0085] In order to make full use of the multi-dimensional information of text features, code value features and statistical features, the present invention designs a multi-expert hybrid model for feature fusion and discrimination. This model can flexibly select and combine the most suitable expert network according to the characteristics of different features to improve the recognition accuracy of key groups.

[0086] The core idea of ​​the multi-expert hybrid model is to assign features to different expert networks, each of which focuses on processing a specific type of feature, such as text features, code value features, or statistical features. Then, the outputs of each expert network are weighted and fused through a gating mechanism to obtain a comprehensive feature representation. This approach can effectively deal with the heterogeneity of multi-dimensional feature data and fully tap the potential information of various features.

[0087] First, the text features, code value features, and statistical features are processed by independent expert networks. Each expert network is a multi-layer perceptron (MLP). In this way, each expert network can independently extract a deep representation of the input features.

[0088] Next, the gating mechanism dynamically adjusts the output weights of each expert network based on the importance of the input features. The gating network generates weighted coefficients for each expert network by analyzing the input features, which determine the contribution of each expert network to the final output.

[0089] Finally, the outputs of all expert networks are fused through weighted summation to generate a comprehensive feature representation. The fused feature representation can express the multi-dimensional characteristics of the group, has stronger discriminative power, and can help the model identify key groups more accurately. The fusion formula is as follows:

[0090]

[0091] Where n is the number of expert networks, F final It is the final feature representation after fusion.

[0092] In addition, the multi-expert hybrid model has strong scalability. It can increase or decrease the number of expert networks according to actual application requirements and optimize the structure of each expert network for specific applications. This makes the present invention widely applicable to different types of social platform data analysis, especially in the face of complex and diverse group information, showing strong robustness and adaptability.

[0093] Key group clues found

[0094] Discovering text clues based on large model prompts is one of the key steps in group analysis. By using pre-trained language models (such as GPT) and designing appropriate prompts, text clues closely related to group behavior can be effectively mined. These clues include the group's charter, emotional tendencies, common topics, hot keywords, and communication styles, helping to identify potential key groups.

[0095] In the application of large language models, the design of prompts directly affects the performance of the model. Prompts are the key to guiding the language model to generate target outputs. They need to be carefully designed according to the specific requirements of the task to ensure that the model can accurately capture key information. In this paper, we designed a series of highly targeted prompts to extract important text clues in groups.

[0096] Here is a prompt design for identifying the core topic of a group:

[0097] "Please analyze the following text and list the three most frequently discussed topics, as well as the possible emotional tendencies behind these topics:\n\n'{Group chat history}'\n\nPlease describe each topic and its related keywords in detail."

[0098] The following clues were obtained from the chat content of China XX website:

[0099] Topics: Internal circulation, employment rate, anti-subsidy

[0100] Emotional tendency: rational and objective

[0101] Keywords: Internal circulation: [international environment, anti-globalization, going overseas], employment rate: [graduation, civil service examination, take-out], anti-subsidy: [new energy vehicles, dairy products, EU],

[0102] In addition, in order to more comprehensively capture the communication style of the group, we also designed Prompt for analyzing conversation characteristics:

[0103] "Please summarize common expressions in the following conversations, including any bias toward specific events:\n\n'{Group chat history}'\n\nAnd indicate whether there are recurring content or specific language patterns."

[0104] Analyzing the comments on China XX website, we get the following clues:

[0105] Is there a preference for a specific event: "Support for independent research and development of electric vehicles"

[0106] By designing and applying these prompts, the present invention can extract highly representative information from the text content of the group and construct a detailed text clue map of the group. These maps include not only the main topics and keywords, but also deep-level features such as emotional tendencies and communication styles, providing a key basis for subsequent group analysis.

[0107] The sliding window time series clue discovery method dynamically analyzes the time dimension of group activities on social platforms to identify specific behaviors. The sliding window technology divides the data into overlapping time periods, and independently analyzes key indicators such as group speech frequency, keyword usage, and user activity in each time period, so as to capture sudden behavioral activities within the group in a timely manner.

[0108] Usually, we set the time granularity of the sliding window to 1 hour or 30 minutes to ensure a quick response to hot events. In each time window, comparing the changes in the current indicator with the previous window can effectively identify the hot time period. For example, a sudden surge in the number of speeches in a certain time window may indicate that the group is focusing on discussing hot events or planning hot activities.

[0109] Taking the speech frequency of China XX website as an example, we get the following clues:

[0110] Suppose we are analyzing the speeches made on China XX website on March 11 and June 12, 2024.

[0111] The data shows that the number of speeches suddenly increased during the following two time periods:

[0112] - 13:00-14:00, March 11, 2024

[0113] -22:00 on June 12, 2024 - 01:00 on June 13, 2024

[0114] The speech time series on March 11 is recorded as X, and the speech time series from June 12 to 13 is recorded as Y. Assume that X = [100, 150, 200, 50] represents the number of speeches per hour from 10:00 to 14:00 on March 11, and Y = [80, 90, 300, 250, 60] represents the number of speeches per hour from 21:00 on June 12 to 1:00 on June 13. Because the lengths of X and Y are different (X has 4 time points and Y has 5 time points), direct comparison may lose some information. At this time, DTW is used to align the two time series and calculate the minimum distance between them to evaluate the similarity of their activity patterns. The calculation formula of DTW is as follows:

[0115]

[0116] Among them, X[i] and Y[j] represent the values ​​of two time series X and Y at the i-th and j-th time points, respectively, and n is the length of the time series. DTW aligns the time series and calculates their similarity by calculating the Euclidean distance of all possible pairs between the two time series and finding the pairing path that minimizes the total distance. The calculation results show the similarity of the speech frequency of China XX network in the two time periods of March 11 and June 12-13.

[0117] Using the DTW results, we can further combine other features (such as sentiment analysis of speech content, keyword extraction, etc.) to perform multi-feature fusion analysis to more accurately determine whether these time periods are high-risk time periods (for example, whether there are potential market fluctuation discussions or sudden policy interpretations). Through these analyses, we can provide early warning information to managers of China XX Network to help them take timely countermeasures when hot events occur.

[0118] Although some embodiments of the present disclosure have been shown and described, it will be appreciated by those skilled in the art that modifications may be made to the embodiments without departing from the principles and spirit of the present disclosure, the scope of which is defined by the claims and their equivalents.

Claims

1. A group clue discovery method based on multi-feature fusion, characterized in that: It consists of two parts: a key group identification module based on multi-feature fusion and a key group clue discovery module; The key group identification module based on multi-feature fusion first uses the GPT model to extract group text data on the social platform, and pre-processes the text data to convert the text data into word embedding representation; Next, the word embedding sequence after position encoding is input into the multi-layer Transformer network to perform deep feature extraction using the self-attention mechanism. The calculation formula is: Where Q represents the query matrix, which indicates the degree of attention of the current word to other words. A query vector is generated for each word position to find related words. K represents the key matrix, which indicates the characteristics or content of the word. A key vector is generated for each word position to match with the query vector (Q) to determine the relevance. V represents the value matrix, which indicates the value or output information of the word. A value vector is generated for each word position to indicate the content information of the word, which will be weighted and summarized into the output. k represents the dimension of the key vector K, which is a normalization factor used to reduce the value range of the dot product result to avoid the saturation of the softmax function due to excessively large values, thereby losing the discrimination. T represents the transposed matrix of the key matrix K. The deep features include code value features, text features, and statistical features. Then, independent expert networks are established for text features, code value features and statistical features. Each of the expert networks is a multi-layer perceptron, which generates representations for text features and reduces the dimension of statistical features. Then, a multi-expert hybrid model is designed to fuse and discriminate the code value features, the representation of text features, and the statistical features after dimension reduction. Specifically, a gating mechanism is designed to generate weighted coefficients for each expert network by analyzing the input features, and dynamically adjust the output weights of each expert network according to the importance of the input features. Finally, the outputs of all expert networks are fused by weighted summation to generate a comprehensive feature representation. The fusion formula is as follows: Where n is the number of expert networks, F final is the final feature representation after fusion; The key group clue discovery module receives the operation results of the multi-expert hybrid model, and screens key groups through text clue discovery based on large model prompts and sliding window time series clue discovery methods, and obtains group screening results.

2. The method for discovering group clues based on multi-feature fusion according to claim 1, characterized in that: The analysis method of the code value feature is: focusing on code values ​​that appear repeatedly in multiple groups or have obvious directionality; The statistical feature analysis method is: using a multi-layer perceptron model to fuse all statistical features to generate the final statistical feature representation. The calculation formula is as follows: F stat =σ(W×X+b) Among them, X is the statistical feature vector of the input, W is the weight matrix, b is the bias, and σ is the activation function.

3. The group clue discovery method based on multi-feature fusion as claimed in claim 2, characterized in that: The text clue discovery based on large model prompts utilizes a pre-trained language model and designs a suitable prompt. When processing the prompt, the model comprehensively considers the word frequency, semantic association and emotional expression in the text to generate an output containing key information, thereby mining text clues closely related to group behavior, including the group's charter, emotional tendencies, common topics, hot keywords and communication style.

4. The method for discovering group clues based on multi-feature fusion as claimed in claim 3, characterized in that: The sliding window time series clue discovery method dynamically analyzes the time dimension of group activities on social platforms to identify specific behaviors. Specifically, the data is divided into overlapping time periods, and the group's speech frequency, keyword usage, and user activity indicators are independently analyzed in each time period to capture sudden behaviors and hot activities within the group. The similarity of the activity patterns of two groups in different time periods is judged. The dynamic time rule method is introduced to measure the time series similarity of the two groups. By calculating the minimum distance between the activity patterns of the groups, the groups with similar activity characteristics in different time periods are identified. The calculation formula is: Among them, X[i] and Y[j] represent the values ​​of the two time series X and Y at the i-th and j-th time points respectively, and n is the length of the time series.

Images