An information analysis method and device, electronic equipment and storage medium
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGZHOU HUYA INFORMATION TECH CO LTD
- Filing Date
- 2026-02-28
- Publication Date
- 2026-06-09
Smart Images
Figure CN122173630A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information analysis and processing, and more specifically, to an information analysis method, apparatus, electronic device, and storage medium. Background Technology
[0002] With the rapid development of live streaming platforms, more and more users are interacting on these platforms. This increase in users means that live streaming platforms need to process a large volume of user-sent messages every day, and the sheer volume and complexity of these messages makes it difficult for platforms to effectively categorize and process them.
[0003] Existing information detection and analysis methods primarily rely on keyword-based rule filtering. However, this approach is limited by predetermined rules and, due to the vast amount of complex user information, cannot cover all possible information categories. Consequently, it fails to effectively analyze user information and hinders the maintenance of live streaming platforms. Therefore, a more effective method for information analysis is urgently needed. Summary of the Invention
[0004] This invention provides an information analysis method, apparatus, electronic device, and storage medium for effectively and efficiently analyzing information.
[0005] According to a first aspect of this application, an information analysis method is provided, the method comprising: Collect the data to be analyzed; Obtain historical data that matches the data to be analyzed from a preset historical knowledge base; A query corpus is constructed based on the matched historical data and the data to be analyzed. The analysis results were obtained by analyzing the query corpus.
[0006] By matching the data to be analyzed with a preset historical knowledge base, historical data matching the data to be analyzed is obtained. This matched historical data is then used as reference information to construct a query corpus together with the data to be analyzed. Furthermore, during the analysis of the query corpus, the information contained in the historical data related to the data to be analyzed is fully integrated, effectively improving the accuracy of the analysis of the data to be analyzed. Optionally, obtaining historical data matching the data to be analyzed from the preset historical knowledge base includes: Candidate historical data are obtained by filtering from the historical data in the historical knowledge base based on the data to be analyzed; Calculate the matching score between the data to be analyzed and the candidate historical data, and obtain the candidate historical data that matches the data to be analyzed from the candidate historical data based on the matching score, as the matched historical data.
[0007] First, candidate historical data are selected based on the data to be analyzed to perform a coarse check of the historical data. Then, by calculating the matching score between each candidate historical data and the data to be analyzed, matching historical data is further selected from the candidate historical data to perform a fine check of the historical data. By combining coarse and fine checks, historical data that is highly correlated with the data to be analyzed can be obtained more accurately.
[0008] Optionally, the step of filtering candidate historical data from historical data in the historical knowledge base based on the data to be analyzed includes: The data to be analyzed and the historical data are encoded respectively to obtain the encoding vectors of the data to be analyzed and the historical data. Calculate the vector similarity between the encoded vector of the data to be analyzed and the encoded vector of the historical data; The historical data is filtered based on the vector similarity to obtain the candidate historical data.
[0009] By encoding the data to be analyzed and each of the historical matching data separately, corresponding encoding vectors are obtained, which can map the data into a unified vector space. This allows for the accurate calculation of the vector similarity between the encoding vector of the data to be analyzed and the encoding vector of the historical data within this vector space. Based on this vector similarity, the historical data is filtered to effectively identify candidate historical data that is semantically or feature-wise closest to the data to be analyzed, thereby significantly improving the accuracy of the matching results. Optionally, calculating the matching score between the data to be analyzed and the candidate historical data includes: The data to be analyzed is concatenated with the candidate historical data to obtain the text to be processed from the candidate historical data; The text to be processed is subjected to relevance processing to obtain the matching score corresponding to the text to be processed.
[0010] By concatenating the data to be analyzed with the candidate historical data to form the text to be processed corresponding to the candidate historical data, and obtaining the corresponding matching score based on the text to be processed, the degree of correlation between the data to be analyzed and the candidate historical data can be effectively quantified, thereby improving the accuracy of the matching results while maintaining the original semantic structure.
[0011] Optionally, the analysis of the query corpus to obtain the analysis results includes: The data to be analyzed is analyzed using the matched historical data to obtain candidate information categories of the data to be analyzed. The candidate information categories are then verified according to preset information category definition rules to determine the final information category of the data to be analyzed. The analysis results are then constructed based on the final information category.
[0012] The data to be analyzed is analyzed using the matched historical data to generate candidate information categories. The candidate information categories are then verified according to preset information category definition rules to ensure that only categories that meet the corresponding information category definition rules are confirmed as the final information categories, thereby improving the accuracy of the information category determination of the data to be analyzed.
[0013] Optionally, the collection of data to be analyzed includes: Collect data to be processed; The data to be processed is matched against a preset target analysis data judgment rule base, and the matched data to be processed is taken as the data to be analyzed.
[0014] By collecting data to be processed and matching the collected data with a preset target analysis data judgment rule library, it is possible to accurately identify data that meets the target characteristics and determine it as the data to be analyzed, thereby effectively improving the accuracy of subsequent analysis and screening.
[0015] Optionally, the method further includes: Extract the information categories of the data to be analyzed based on the analysis results; A management decision text is generated based on the information category of the data to be analyzed, and the data to be analyzed is processed based on the management decision text.
[0016] Based on the analysis results, information categories of the data to be analyzed are extracted, and management decision texts are generated accordingly. Then, the data to be analyzed is processed based on the management decision texts, which can achieve precise and standardized handling of the data to be analyzed, and significantly improve the pertinence and execution effect of the processing operations.
[0017] According to a second aspect of this application, an information analysis apparatus is provided, the apparatus comprising: The data acquisition module is used to collect the data to be analyzed. The data matching module is used to obtain historical data that matches the data to be analyzed from a preset historical knowledge base; The corpus construction module is used to construct a query corpus based on the matched historical data and the data to be analyzed; The information analysis module is used to analyze the query corpus and obtain analysis results.
[0018] According to a third aspect of this application, an electronic device is provided, comprising: Memory, used to store one or more computer programs; A processor, when the one or more computer programs are executed by the processor, implements the information analysis method described in the first aspect above.
[0019] According to a fourth aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the information analysis method described in the first aspect above.
[0020] Based on any of the above aspects, the information analysis method, apparatus, electronic device, and computer storage medium provided in this application embodiment match the data to be analyzed with a preset historical knowledge base to obtain historical data that matches the data to be analyzed, and use the matched historical data as reference information to jointly construct an inquiry corpus with the data to be analyzed. In the process of analyzing the inquiry corpus, the information contained in the historical data that is related to the data to be analyzed is fully integrated, thereby effectively improving the accuracy of the analysis of the data to be analyzed. Attached Figure Description
[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0022] Figure 1 This is a schematic application scenario diagram of the information analysis method provided in this embodiment.
[0023] Figure 2 A flowchart of the information analysis method provided in this embodiment.
[0024] Figure 3 This is a flowchart illustrating the steps involved in acquiring historical data in this embodiment.
[0025] Figure 4 This is a flowchart illustrating the steps for obtaining candidate historical data in this embodiment.
[0026] Figure 5 This is a schematic diagram of the steps for obtaining the matching score provided in this embodiment.
[0027] Figure 6 This is a schematic diagram of the functional modules of the information analysis device provided in this embodiment.
[0028] Figure 7 This embodiment provides a schematic diagram of the electronic device. Detailed Implementation
[0029] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this application. To better illustrate the following embodiments, some components in the drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions of the product; it is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.
[0030] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.
[0031] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0032] With the rapid development of live streaming platforms, more and more users are interacting on these platforms. This increase in users means that live streaming platforms need to process a large volume of user-sent messages every day, and the sheer volume and complexity of these messages makes it difficult for platforms to effectively categorize and process them.
[0033] This information includes black market information, which refers to illegal information sent by black market users. Black market users use live streaming platforms to send illegal black market information to other users to obtain profits. Due to the large number of users on the platform, it is very difficult to analyze and filter out black market information from all the information on the live streaming platform.
[0034] Existing information detection and analysis methods are mainly based on keyword rule filtering or deep learning. These methods can analyze information on live streaming platforms to a certain extent and filter out black market information. However, keyword rule filtering is limited by established rules and cannot cover all information categories, thus failing to effectively identify black market information. Deep learning requires a large amount of labeled data for training and is not effective in judging complex texts. As a result, existing information analysis techniques are unable to fully achieve effective analysis of different types of information.
[0035] This embodiment provides a technical solution that can solve the above problems. The specific implementation of this application will be described in detail below with reference to the accompanying drawings.
[0036] This is an exemplary schematic diagram illustrating an application scenario of an information analysis method provided in an embodiment of this application. Figure 1 As shown, the application scenario includes at least a server 100 and a terminal 200 that can communicate with the server 100.
[0037] Understandably, the server 100 can be an independent electronic device or a cluster of multiple electronic devices; the terminal 200 can be a smartphone terminal, personal computer, tablet computer, vehicle terminal, etc., but is not limited to these.
[0038] In one feasible approach, server 100 and terminal 200 may each execute an information analysis method provided in the embodiments of this application. Alternatively, the information analysis method provided in the embodiments of this application may be executed partly in server 100 and partly in terminal 200.
[0039] like Figure 2 As shown, this embodiment provides an information analysis method, which may include the following steps: S1: Collect the data to be analyzed; In this embodiment, the collection of information data may include the following steps: Collect data to be processed; In one implementation, the data to be processed can be text data from a live streaming platform. This text data may include user private messages, user profiles, usernames, and user comments, etc. Simultaneously, the user corresponding to the data is recorded during the data collection process. By collecting various text data from the live streaming platform, relevant information about all users on the platform can be extracted, including information related to illegal activities.
[0040] The data to be processed is matched against a preset target analysis data judgment rule base, and the matched data to be processed is taken as the data to be analyzed.
[0041] In this embodiment, the target analysis data judgment rule library contains several preset target analysis data judgment rules. These target analysis data judgment rules are obtained by analyzing and summarizing all text data on the live streaming platform. They are used to filter the data to be processed, remove invalid data, and effectively filter the data to be processed to identify the information categories that need to be further analyzed in this embodiment, thereby reducing the amount of data that needs to be processed subsequently.
[0042] Understandably, in this embodiment, the target analysis data judgment rules in the target analysis data judgment rule base can be used to filter out the data to be processed that is suspected of containing black market information, so that the information analysis method described in this embodiment can analyze the data to be analyzed, filter out the black market information, and realize the supervision of the live streaming platform.
[0043] Therefore, the target analysis data judgment rule base can include the following judgment rules for screening black market information.
[0044] Duplicate text sending: If a message sender sends the same text message more than the preset number of times within a preset first time range, it is determined to be duplicate text sending. Single-address user private message overload: If a user at a single address sends private messages to more than a preset number of users at a second time within a preset time range, it is determined as single-address user private message overload. User private message overload per unit time: Set corresponding private message threshold limits for different time ranges. Within each time range, if the number of private messages sent by a user exceeds the corresponding private message threshold limit, such as a private message threshold limit of 5 for 10 minutes, 10 for 30 minutes, and 20 for 60 minutes, then if the number of private messages exceeds 5 within 10 minutes, 10 within 30 minutes, or 20 within 60 minutes, it can be determined that the user has exceeded the private message overload per unit time.
[0045] By using the aforementioned target analysis data determination rules, data suspected of containing black market information can be effectively filtered from the data to be processed. It should be noted that the determination rules shown above are merely examples and do not include all the determination rules in the target analysis data determination rule library. The target analysis data determination rules can be set based on the actual text data and are not limited to the types shown above. It is understood that in actual use, a single target analysis data determination rule can be matched against the data to be processed, or multiple target analysis data determination rules can be combined to form a determination rule group for matching against the data to be processed.
[0046] In this embodiment, the matching of the data to be processed can be achieved through a rule engine. The target analysis data determination rule library is set in the rule engine. The rule engine can use the determination rules or determination rule groups in the target analysis data determination rule library to query the data to be processed and determine whether there is a target analysis data determination rule or determination rule group that matches the data to be processed. If there is a target analysis data determination rule or determination rule group that matches the data to be processed, the data to be processed is determined to be data to be analyzed; otherwise, the data to be processed is determined to be redundant information data.
[0047] S2: Obtain historical data that matches the data to be analyzed from a preset historical knowledge base; In this embodiment, the construction of the historical knowledge base may include: Collect historical text samples with known information categories, and label the historical text samples according to the information categories to obtain historical data containing historical text samples and information categories.
[0048] In one implementation, a maximum text length can be set for each historical text sample contained in the historical data. The maximum text length can be set according to the maximum processing length of the model required for subsequent processing. By setting the maximum text length, it is possible to ensure that the input is not truncated during the model processing.
[0049] In one implementation, the historical data in the historical knowledge base are independent of each other, that is, the historical data is stored in blocks. By storing the historical data in blocks, noise can be reduced when retrieving the historical data and the accuracy of matching the historical data can be improved.
[0050] In this embodiment, as Figure 3 As shown, step S2 may include the following steps: S21: Based on the data to be analyzed, candidate historical data are obtained by filtering from the historical data in the historical knowledge base; S22: Calculate the matching score between the data to be analyzed and the candidate historical data, and obtain the candidate historical data that matches the data to be analyzed from the candidate historical data according to the matching score, as the matched historical data.
[0051] Furthermore, such as Figure 4 As shown, step S21 may include the following sub-steps: S211: Encode the data to be analyzed and the historical data respectively to obtain the encoding vectors of the data to be analyzed and the historical data; S212: Calculate the vector similarity between the encoded vector of the data to be analyzed and the encoded vector of the historical data; S213: Filter the historical data based on the vector similarity to obtain the candidate historical data.
[0052] In this embodiment, the historical data in the historical knowledge base can be processed using a pre-trained vector coding model to obtain the coding vector of each historical data. The data to be analyzed can also be processed using the vector coding model to obtain the coding vector of the data to be analyzed. The vector coding model can be based on an existing model with coding capabilities or based on an embedded encoder. The coding vector of the data to be analyzed is then compared with the coding vector of each historical data to obtain the vector similarity corresponding to each historical data.
[0053] After obtaining the vector similarity corresponding to each of the historical data, the historical data are sorted in descending order of the vector similarity. Based on the sorted historical data, a number of historical data with the highest vector similarity are selected as candidate historical data.
[0054] In this embodiment, as Figure 5 As shown, calculating the matching score between the data to be analyzed and the candidate historical data in step S22 may include the following sub-steps: S221: Concatenate the data to be analyzed with the candidate historical data to obtain the text to be processed from the candidate historical data; S222: Perform relevance processing on the text to be processed to obtain the matching score of the text to be processed.
[0055] In this embodiment, the candidate historical data can be concatenated with each of the data to be analyzed using preset marker symbols to form a text to be processed containing data pairs of data to be analyzed and candidate historical data, such as "data to be analyzed [SEP] candidate historical data", where "SEP" represents a separator marker symbol. The concatenated text to be processed is then subjected to correlation processing, which involves learning and understanding the data to be analyzed and the candidate historical data in the input text to obtain the correlation between the data to be analyzed in the text, and scoring is performed based on the correlation.
[0056] In this embodiment, step S222 can be as follows: perform relevance processing on the text to be processed using a pre-trained text relevance detection model to obtain a matching score corresponding to the text to be processed. The text relevance detection model can perform semantic analysis on the data to be analyzed and the candidate historical data in the input text to be processed, and obtain the semantic relevance between the data to be analyzed and the candidate historical data. Based on the semantic relevance, a score is given to obtain the matching score.
[0057] In one implementation, the text relevance detection model can be built based on a semantic reordering model, which includes a joint encoding module, an attention interaction module, and an output scoring module connected in sequence. The joint encoding module converts the input text into a unified input representation that the model can process. The attention interaction model contains several Transformer units, and the attention interaction module uses the self-attention mechanism of the Transformer units to generate semantic association features between the data to be analyzed and the candidate historical data in the text to be processed. The output scoring model predicts the relevance score between the data to be analyzed and the candidate historical data in the text to be processed based on the semantic association features.
[0058] In one implementation, after obtaining the matching score, candidate historical data whose matching scores exceed a preset matching score threshold can be used as the historical matching data. It is understood that if there is no candidate historical data exceeding the matching score threshold, the historical matching data is empty.
[0059] Compared to calculating similarity, obtaining a matching score through the text relevance detection model can more accurately determine the similarity relationship between the data to be analyzed and the candidate historical data. Therefore, step S21 can be understood as performing a coarse search on the data to be analyzed, and step S22 can be understood as performing a fine search on the data to be analyzed. The coarse search can reduce the amount of data to be processed by the fine search and improve the efficiency of data processing. The two-level search can ensure that the final historical data and the data to be analyzed have a high degree of relevance, reducing the interference of the subsequent information analysis model's reasoning process.
[0060] S3: Construct a query corpus based on the matched historical data and the data to be analyzed; In this embodiment, the query corpus can be constructed using a role-based prompting architecture. First, the identity of the processing model is defined, constraining its task responsibilities and output format requirements. Then, specific analysis tasks are provided. These analysis tasks can be constructed based on the matched historical data and the data to be analyzed.
[0061] In one implementation, the model's task can be set to analyze the information category of the data to be analyzed, using the matched historical data as a reference. By constructing the query corpus using a role-based prompting architecture, the model can be effectively guided to perform the corresponding task, determining the information category corresponding to the data to be analyzed based on historical similar data.
[0062] S4: Analyze the query corpus to obtain the analysis results.
[0063] In this embodiment, step S4 can be performed by analyzing the query corpus using a pre-trained information analysis model to obtain analysis results. The information analysis model can be an existing question-answering model or language processing model with logical reasoning capabilities. This model can fully understand the text content input into the model and perform logical reasoning based on that content to obtain inference results. Therefore, after the query corpus is input into the model, the description of the analysis task set in the corpus will guide the model, and the model's logical reasoning capabilities will be used to execute the corresponding analysis task.
[0064] Since the analysis task set in the query corpus is based on the historical matching data as a reference to determine the information category corresponding to the data to be analyzed, in one embodiment, the analysis of the query corpus to obtain the analysis result in step S4 may include: In the information analysis model, the data to be analyzed is analyzed using the matched historical data to obtain the information category to which the data to be analyzed belongs, and the analysis result is constructed based on the information category to which the data to be analyzed belongs.
[0065] In this embodiment, the information category can be divided into illegal black market information category and normal information category; wherein, the analysis can include, based on its logical reasoning ability, the information analysis model compares the semantic relevance of the data to be analyzed with the semantics of each of the historical data, and determines the information category to which the data to be analyzed belongs based on the information category marked in the matched historical data, that is, whether the data to be analyzed belongs to the black market information category or the normal information category.
[0066] As mentioned above, since the historical matching data may be empty, the information analysis model lacks a reference basis for processing. Therefore, in one embodiment, the query corpus may further include preset information category definition rules. These rules may include a category description of the information category and a classification rule description corresponding to that category description. The category description is a textual description of the information category. In one example of this embodiment, the category description may include textual descriptions of black market information categories such as vulgar traffic generation, illegal advertising traffic generation, and vulgar content, as well as textual descriptions of normal information categories. The classification rule description serves as the judgment condition for the corresponding information category, assisting the information analysis model in determining the information category of the data to be analyzed. It is understood that since there are multiple information categories, the query corpus may contain information category definition rules for all of these information categories, with each information category corresponding to one information category definition rule, to facilitate the information analysis model's judgment.
[0067] In this embodiment, the step S4 of analyzing the query corpus to obtain the analysis result may include: In the information analysis model, the data to be analyzed is analyzed using the matched historical data to obtain candidate information categories of the data to be analyzed. The candidate information categories are then verified according to the information category definition rules to determine the final information category of the data to be analyzed. The analysis results are then constructed based on the final information category.
[0068] Understandably, in this embodiment, the candidate information categories can be divided into black market information categories and normal information categories. Similarly, the final information category can include black market information categories and normal information categories. Understandably, in the information analysis model, firstly, the correlation between the historical data of each match and the data to be analyzed is compared to determine the candidate information categories of the data to be analyzed. Then, based on the candidate information categories, the data to be analyzed is compared with the information category definition rules corresponding to the candidate information categories to verify whether the candidate information categories are accurate. If the candidate information categories are accurate, then the candidate information categories are used as the final information categories. If the candidate information categories are inaccurate, then the data to be analyzed is sequentially compared with the preset information category definition rules for each information category.
[0069] Understandably, if the historical data is empty, the information analysis model directly determines the data to be analyzed based on the information category definition rules, determines the final information category of the data to be analyzed, and constructs the analysis results.
[0070] Understandably, after obtaining the final information category and constructing the analysis result, if the final information category in the analysis result is a black market information category, then the data to be analyzed can be determined to be black market information data; conversely, if the final information category is a normal information category, then the data to be analyzed can be determined to be normal information data.
[0071] To facilitate the acquisition of different labels and the generation of different automatic decision-making methods, in one embodiment, the query corpus also includes an output structure format, which the information analysis model outputs according to the structure format to obtain analysis results conforming to the structure format. The structure format includes an analysis approach and analysis results; the analysis approach is used to visualize the analysis process of the information analysis model, making the analysis results interpretable.
[0072] In one example, the analysis result can be: { Analysis Approach: 1. The historical data categorizes this as "Violation of Traffic Diversion," but the current data being analyzed does not contain any explicit traffic diversion entities such as pure URLs or communication links, thus failing to meet the definition rules for the "Violation of Traffic Diversion" information category. 2. The phrase "Customized Dance, Super Seductive" in the data being analyzed meets the definition rules for "Vulgar Content." 3. Although the historical data categorizes this as "Violation of Traffic Diversion," the data being analyzed does not meet the conditions of the definition rules for the "Violation of Traffic Diversion" information category, therefore it does not constitute a "Violation of Traffic Diversion." 4. Based on the overall assessment, it is determined to be "Vulgar Content." Analysis Results: The information category of the data to be analyzed is "vulgar content".} In some implementations, step S4 may further involve mapping the entities, attributes, and relationships in the query corpus to a preset knowledge graph to form a subgraph to be analyzed; performing structured matching and path reasoning based on the subgraph to be analyzed and existing historical fact subgraphs in the knowledge graph; and determining the analysis result of the data to be analyzed based on the matching result and the reasoning path.
[0073] By mapping the query corpus to a knowledge graph and performing structured matching and path reasoning, the semantic association between the data to be analyzed and the matched historical data can be explicitly depicted using the predefined entity relationships and logical structures in the knowledge graph, thereby achieving more accurate and interpretable analysis results under a unified knowledge system.
[0074] In one embodiment, the method may further include: Based on the analysis results, extract the information categories of the data to be analyzed; generate management decision text based on the information categories of the data to be analyzed, and process the data to be analyzed based on the management decision text.
[0075] It is understandable that if the data to be analyzed is normal information data, then no processing will be performed on the data to be analyzed.
[0076] In this embodiment, corresponding management decision texts can be pre-set for different information categories. The management decision texts contain processing methods for the data to be analyzed for the corresponding information category. For example, for data to be analyzed involving vulgar content, the sender of the data to be analyzed can be prohibited from sending text messages for a certain period of time. For data to be analyzed that violates traffic diversion regulations, the sender of the data to be analyzed can be permanently prohibited from sending text messages.
[0077] Understandably, if the data to be analyzed is black market information, the historical knowledge base can be updated based on the data to be analyzed and the information category corresponding to the data to be analyzed, thereby continuously enriching the historical data of the historical knowledge base and providing sufficient basis for subsequent matching analysis of the data to be analyzed.
[0078] In this embodiment, the data to be processed is first filtered using a set target analysis data judgment rule base. This reduces the amount of data to be processed and extracts the data that needs to be analyzed, thus reducing misjudgments. By constructing a historical knowledge base, query corpus is built using historical data that matches the data to be analyzed and the data to be analyzed. The logical reasoning ability of the information analysis model is used to perform semantic-level correlation analysis on the historical data and the data to be analyzed. This allows for effective analysis and identification of the information categories in the data to be analyzed without being restricted by predetermined rules, thereby enabling effective monitoring of the data to be analyzed based on the information categories.
[0079] like Figure 6 As shown in the illustration, this application also provides an information analysis device. Optionally, the information analysis device may include: Data acquisition module 11 is used to acquire data to be analyzed; In this embodiment, the data acquisition module 11 can be used to perform... Figure 2 For a detailed description of the data acquisition module 11 shown in step S1, please refer to the description of step S1.
[0080] Data matching module 12 is used to obtain historical data that matches the data to be analyzed from a preset historical knowledge base; In this embodiment, the data matching module 12 can be used to perform... Figure 2 For a detailed description of the data matching module 12 shown in step S2, please refer to the description of step S2.
[0081] Corpus construction module 13 is used to construct query corpus based on the matched historical data and the data to be analyzed; In this embodiment, the corpus construction module 13 can be used to perform... Figure 2 For a detailed description of the corpus construction module 13 shown in step S3, please refer to the description of step S3.
[0082] Information analysis module 14 is used to analyze the query corpus to obtain analysis results; In this embodiment, the violation detection module 14 can be used to perform... Figure 2 For a detailed description of the violation detection module 14 shown in step S4, please refer to the description of step S4.
[0083] It is understood that the above-described device embodiments and method embodiments can correspond to each other, and similar descriptions of the device embodiments can be referred to the method embodiments. To avoid repetition, further details are omitted here. The information analysis device provided in this application can execute an information analysis method provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects for executing the method. The functional modules of the information analysis device can be implemented in hardware, in software instructions, or in a combination of hardware and software modules.
[0084] Specifically, the steps of the method embodiments of this application can be implemented by integrated logic circuits in the processor hardware and / or instructions in software form. The steps of the information analysis method in conjunction with the embodiments of this application can be directly manifested as execution by a hardware encoding processor, or execution by a combination of hardware and software modules in the encoding processor. Optionally, the software module can be located in random access memory, and storage media such as read-only memory, programmable read-only memory, flash memory, electrically erasable programmable memory, and registers are all acceptable. The storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps in the above method embodiments.
[0085] This application provides an electronic device with the following structure: Figure 7 As shown. The electronic device can be as described in this embodiment. Figure 1 The server 100 or terminal 200 shown.
[0086] The electronic device includes a memory 21, a processor 22, a communication module 23, and an input / output interface 24, etc. Optionally, the memory 21, the processor 22, the communication module 23, and the input / output interface 24 can be connected and communicate with each other through a bus 25.
[0087] The memory 21 is used to store one or more computer programs and to transfer the code of the computer programs to the processor 22; when the one or more computer programs are executed by the processor 22, the information analysis method in the embodiments of this application is implemented.
[0088] Optionally, the electronic device can be connected to a network via communication module 23 to communicate with other devices, such as terminals or servers, to achieve data interaction. The electronic device can be various forms of digital computers, exemplarily such as desktop computers, servers, workbenches, mainframes, or other types of computers. The electronic device can also be various forms of mobile terminals, exemplarily such as smartphones, tablets, wearable devices (such as helmets, glasses, watches, etc.), and other similar mobile terminals.
[0089] Optionally, the electronic device can connect to required input / output devices, such as a keyboard or display device, via the input / output interface 24. The electronic device itself may have a display device, and other display devices can also be connected externally via the input / output interface 24. Optionally, a storage device, such as a hard disk, can also be connected via the input / output interface 24 to store data from the electronic device, read data from the storage device, or store data from the storage device in the memory 21. It is understood that the input / output interface 24 can be a wired interface or a wireless interface. Depending on the actual application scenario, the device connected to the input / output interface 24 can be a component of the electronic device or an external device connected to the electronic device when needed.
[0090] Optionally, the memory 21 may be a volatile memory and / or a non-volatile memory. The volatile memory may be a random access memory, etc., and the non-volatile memory may be a read-only memory, a programmable read-only memory, an erasable programmable read-only memory, an electrically erasable programmable read-only memory, or a flash memory, etc.
[0091] Optionally, the computer program stored in the processor 22 can be divided into one or more modules, which are stored in the memory 21 and executed by the processor 22 to perform the method provided in this embodiment. The one or more modules can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the electronic device.
[0092] Optionally, the processor 22 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 22 include, but are not limited to, a central processing unit, a graphics processing unit, a digital signal processor, various special-purpose artificial intelligence computing chips, various processors running machine learning model algorithms, and can also be any suitable controller, microcontroller, processor, etc. The processor 22 executes the various methods and processes of this embodiment, exemplarily, such as an information analysis method according to an embodiment of this application.
[0093] Optionally, the bus 25 may include a path for transmitting information. Depending on its function, the bus 25 may be divided into an address bus, a data bus, a control bus, etc.
[0094] In an optional implementation, this application embodiment also provides a computer storage medium storing a computer program thereon. When executed by a computer, the computer program enables the computer to perform the methods described in the above-described method embodiments. Part or all of the computer program can be loaded and / or installed on the memory 21 of an electronic device. When the computer program is executed by the processor 22, one or more steps of an information analysis method according to an embodiment of this application can be performed.
[0095] Optionally, the computer-readable storage medium may be a random access memory, a read-only memory, a programmable read-only memory, an erasable programmable read-only memory, an electrically erasable programmable read-only memory, etc.
[0096] Obviously, the above embodiments of this application are merely examples for clearly illustrating the technical solution of this application, and are not intended to limit the specific implementation of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the claims of this application should be included within the protection scope of the claims of this application.
Claims
1. An information analysis method, characterized in that, The method includes: Collect the data to be analyzed; Obtain historical data that matches the data to be analyzed from a preset historical knowledge base; A query corpus is constructed based on the matched historical data and the data to be analyzed. The analysis results were obtained by analyzing the query corpus.
2. The information analysis method according to claim 1, characterized in that, The step of obtaining historical data matching the data to be analyzed from a preset historical knowledge base includes: Candidate historical data are obtained by filtering from the historical data in the historical knowledge base based on the data to be analyzed; Calculate the matching score between the data to be analyzed and the candidate historical data, and obtain the candidate historical data that matches the data to be analyzed from the candidate historical data based on the matching score, as the matched historical data.
3. The information analysis method according to claim 2, characterized in that, The step of filtering candidate historical data from historical data in the historical knowledge base based on the data to be analyzed includes: The data to be analyzed and the historical data are encoded respectively to obtain the encoding vectors of the data to be analyzed and the historical data. Calculate the vector similarity between the encoded vector of the data to be analyzed and the encoded vector of the historical data; The historical data is filtered based on the vector similarity to obtain the candidate historical data.
4. The information analysis method according to claim 2, characterized in that, The calculation of the matching score between the data to be analyzed and the candidate historical data includes: The data to be analyzed is concatenated with the candidate historical data to obtain the text to be processed from the candidate historical data; The text to be processed is subjected to relevance processing to obtain a matching score.
5. The information analysis method according to claim 1, characterized in that, The analysis results obtained by analyzing the query corpus include: The data to be analyzed is analyzed using the matched historical data to obtain candidate information categories of the data to be analyzed. The candidate information categories are then verified according to preset information category definition rules to determine the final information category of the data to be analyzed. The analysis results are then constructed based on the final information category.
6. An information analysis method according to any one of claims 1-5, characterized in that, The collected data to be analyzed includes: Collect data to be processed; The data to be processed is matched against a preset target analysis data judgment rule base, and the matched data to be processed is taken as the data to be analyzed.
7. An information analysis method according to any one of claims 1-5, characterized in that, The method further includes: Extract the information categories of the data to be analyzed based on the analysis results; A management decision text is generated based on the information category of the data to be analyzed, and the data to be analyzed is processed based on the management decision text.
8. An information analysis device, characterized in that, The device includes: The data acquisition module is used to collect the data to be analyzed. The data matching module is used to obtain historical data that matches the data to be analyzed from a preset historical knowledge base; The corpus construction module is used to construct a query corpus based on the matched historical data and the data to be analyzed; The information analysis module is used to analyze the query corpus and obtain analysis results.
9. An electronic device, characterized in that, include: Memory, used to store one or more computer programs; A processor, when the one or more computer programs are executed by the processor, implements an information analysis method as described in any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute an information analysis method as described in any one of claims 1-7.