Intelligent data report generation method based on multi-dimensional feature mining and related equipment

Abstract

The invention provides a multi-dimensional feature mining-based data report intelligent generation method and related equipment. The method comprises the following steps of: obtaining clue data to be processed; and determining to-be-extracted entities and attributes in the clue data according to pre-extracted core analysis elements, and establishing an association relationship between the entities. And constructing a first cue word according to the entity, the attribute, the association relationship and the clue data, inputting the first cue word into a large language model, and outputting key data corresponding to the clue data through the large language model. And performing statistical analysis on the key data by using a statistical algorithm to obtain a statistical analysis result. And updating a pre-constructed first analysis template according to the statistical analysis result to obtain a second analysis template. And filling the second analysis template based on the key data to generate the analysis report. And the analysis report is automatically generated, so that the generation rate and accuracy of the analysis report can be improved, and the requirement of a rapidly changing decision-making environment is met.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method for intelligently generating data reports based on multi-dimensional feature mining and related equipment. Background Art

[0002] With the advent of the information age, the amount and complexity of data are increasing, especially in the fields of social governance and law. Facing a large number of events and related clues, traditional manual processing and analysis methods can no longer meet the needs of rapid response and processing. In particular, when generating analysis reports for specific political and legal clues, traditional methods are often inefficient, prone to errors, and difficult to adapt to the rapid changes in data characteristics.

[0003] In addition, manual report preparation is not only time-consuming and labor-intensive, but also difficult to ensure the timeliness and depth of the report content. With the surge in data volume, extracting key information from this data and generating insightful reports has become a major challenge. These reports usually need to reflect the latest data insights and trends to support decision-making, and manual methods are often unable to cope with large amounts of complex data. Summary of the invention

[0004] In view of this, the purpose of this application is to propose a data report intelligent generation method and related equipment based on multidimensional feature mining to solve the problem of low efficiency and poor timeliness of analysis report generation.

[0005] Based on the above objectives, the first aspect of the present application provides a method for intelligently generating data reports based on multidimensional feature mining, comprising:

[0006] Get the clue data to be processed;

[0007] Determine entities and attributes to be extracted in the clue data according to the pre-extracted core analysis elements, and establish association relationships between entities; wherein the core analysis elements are determined according to the historical report data;

[0008] Construct a first prompt word according to the entity, the attribute, the association relationship and the clue data, input the first prompt word into a large language model, and output key data corresponding to the clue data through the large language model; wherein the first prompt word is used to instruct the large language model to extract key data from the clue data according to the entity, the attribute and the association relationship;

[0009] Performing statistical analysis on the key data using a statistical algorithm to obtain statistical analysis results;

[0010] According to the statistical analysis result, the pre-constructed first analysis template is updated to obtain a second analysis template; wherein the first analysis template is determined according to the historical report data;

[0011] The second analysis template is filled based on the key data to generate the analysis report.

[0012] The second aspect of the present application provides a data report intelligent generation device based on multi-dimensional feature mining, comprising:

[0013] An acquisition module, configured to acquire clue data to be processed;

[0014] A determination module, configured to determine entities and attributes to be extracted in the clue data according to the pre-extracted core analysis elements, and to establish association relationships between entities; wherein the core analysis elements are determined according to the historical report data;

[0015] an extraction module, configured to construct a first prompt word according to the entity, the attribute, the association relationship and the clue data, input the first prompt word into a large language model, and output key data corresponding to the clue data through the large language model; wherein the first prompt word is used to instruct the large language model to extract key data from the clue data according to the entity, the attribute and the association relationship;

[0016] An analysis module is configured to perform statistical analysis on the key data using a statistical algorithm to obtain a statistical analysis result;

[0017] An updating module, configured to update a pre-built first analysis template according to the statistical analysis result to obtain a second analysis template; wherein the first analysis template is determined according to historical report data;

[0018] A generation module is configured to fill in the second analysis template based on the key data to generate the analysis report.

[0019] The third aspect of the present application also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the method described in the first aspect when executing the computer program.

[0020] The present application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, wherein the computer instructions are used to enable a computer to execute the method described in the first aspect.

[0021] From the above, it can be seen that the intelligent data report generation method and related equipment based on multidimensional feature mining provided by the present application include obtaining clue data to be processed; determining the entities and attributes to be extracted in the clue data according to the pre-extracted core analysis elements, and establishing the association relationship between the entities; wherein the core analysis elements are determined according to the historical report data, so as to accurately determine the specific content to be extracted from the clue data. A first prompt word is constructed according to the entity, the attribute, the association relationship and the clue data, and the first prompt word is input into the large language model, and the key data corresponding to the clue data is output through the large language model; wherein the first prompt word is used to instruct the large language model to extract key data from the clue data according to the entity, the attribute and the association relationship. Using the understanding ability of the large language model, the large language model is instructed to extract key data through the first prompt word to provide a data basis for the subsequent generation of an analysis report. The key data is statistically analyzed using a statistical algorithm to mine new trends and new content in the key data to obtain statistical analysis results. The first analysis template constructed in advance is updated according to the statistical analysis result to obtain the second analysis template; wherein the first analysis template is determined according to the historical report data; and the dynamic adjustment of the analysis template is realized to adapt to the real-time change trend of the clue data. The second analysis template is filled based on the key data to generate the analysis report. Automatically generating the analysis report helps to improve the generation rate and accuracy of the analysis report, ensure that the report content is updated as the data is updated, and meet the needs of the rapidly changing decision-making environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the present application or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 A schematic diagram of a flow chart of a method for intelligently generating data reports based on multidimensional feature mining according to an embodiment of the present application;

[0024] Figure 2 A flowchart of a method for determining core analysis elements of an embodiment of the present application;

[0025] Figure 3 This is a schematic diagram of the structure of a device for intelligently generating data reports based on multi-dimensional feature mining according to an embodiment of the present application;

[0026] Figure 4 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0028] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be the usual meanings understood by people with ordinary skills in the field to which the present application belongs. The "first", "second" and similar words used in the embodiments of the present application do not represent any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing in front of the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0029] As described in the background technology, when generating analysis reports of political and legal clues, there are problems of low efficiency and poor timeliness. Based on the above problems, this application proposes a method for intelligently generating data reports based on multidimensional feature mining, which can effectively extract key information from a large amount of data, conduct in-depth statistical analysis, and dynamically adjust the report content according to the analysis results. The method of this application can not only improve the efficiency and accuracy of the generation of analysis reports, but also ensure that the report content is updated as the data is updated, thereby truly meeting the needs of the rapidly changing decision-making environment.

[0030] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0031] Figure 1 A method for intelligently generating data reports based on multidimensional feature mining is shown in Figure 1 , including the following steps:

[0032] Step 102: Obtain clue data to be processed.

[0033] Specifically, clue data is unstructured data, which is data described in natural language. Clue data can include various types of data. Taking data in the political and legal fields as an example, clue data can include real estate-related data, disaster-related data, and traffic accident data, etc. These clue data are usually stored in databases, information systems, or information platforms, which may include internal databases, public records, and other related information systems. When obtaining clue data, it is necessary to ensure that you have data access rights and obtain clue data under the premise of complying with data protection regulations.

[0034] Step 104: determine the entities and attributes to be extracted in the clue data based on the pre-extracted core analysis elements, and establish association relationships between entities; wherein the core analysis elements are determined based on historical report data.

[0035] Specifically, the core analysis elements are determined based on the historical report data. The core analysis elements reflect the core content contained in the historical report data, and are also the content that needs to be reflected in the analysis report corresponding to the clue data. The entities and attributes involved in the business process corresponding to the clue data are determined based on the core analysis elements. The core analysis elements may specifically include clue data type, time, location, participant information, and number of participants, etc. For example, in the scenario of housing-related event analysis, entities may include "real estate developers", "event participants", "time" and "location", etc., among which the attributes of "event participants" include "name", "gender" and "contact information", etc.

[0036] Define the relationship between entities according to the business logic corresponding to the clue data. The relationship between entities is crucial for data analysis and report generation. It can help identify the internal correlation of clue data and better complete data mining and data analysis. For example, in the scenario of real estate-related event analysis, possible relationships include "event and participant", "event and facility property", etc.

[0037] Determined entities and attributes can be stored in the database through a pre-built data model. When the historical report data changes, the entities and attributes stored in the database will also change accordingly. Data model building is the process of designing and defining how to organize, store and manage data in a database or data platform. Data model building aims to ensure that the data structure can support business logic and analysis requirements while maintaining the scalability, security and maintainability of the data. The data model also sets data integrity rules, such as primary keys, foreign keys and other business rules to ensure data accuracy and consistency. The primary key is a field or combination of fields that is uniquely used to identify records in a database table. Each table can only have one primary key, and the value of the primary key is unique and cannot be null. A foreign key is a field in a table whose value refers to the primary key in another table. Users establish and maintain the relationship between the two tables to ensure data consistency and integrity. In the database, the ID of each clue data is unique. The definition of relationships between entities not only helps understand how data is related, but also supports database operations such as join queries.

[0038] At the same time, it is also necessary to determine the security and access mechanism of the database, determine the access rights and access operations that each user can perform, and protect sensitive data in the database from being accessed by unauthorized users. In addition, data security measures can be implemented, such as encrypting sensitive fields, to comply with data protection regulations and maintain user trust.

[0039] Step 106: construct a first prompt word according to the entity, the attribute, the association relationship and the clue data, input the first prompt word into a large language model, and output key data corresponding to the clue data through the large language model; wherein the first prompt word is used to instruct the large language model to extract key data from the clue data according to the entity, the attribute and the association relationship.

[0040] Specifically, after the entities, attributes and associations are determined, the key data in the clue data can be extracted through the large language model. Key data is data used to generate analysis reports. In this embodiment, the large language model is used to process the data described in natural language (clue data) and extract the key data. Specifically, a first prompt word is constructed based on the entities, attributes and associations so that the large language model understands the specific content to be extracted, and extracts the key data in the clue data according to the first prompt word, for example, extracting key information such as event type, time and place of event occurrence, etc.

[0041] After the key data is extracted, it is necessary to verify the key data by checking samples or automated testing to verify whether the extracted key data accurately reflects the core analysis elements. Furthermore, data verification includes random sampling checks of key data and comparison verification with annotated data. The data verification results are used to feed back and optimize the large language model to ensure the accuracy and efficiency of key data extraction.

[0042] The extracted key data can be converted into structured data and stored to form a data set that is easy to analyze. Each key data exists in a formatted manner, such as JSON format, which contains detailed information such as event type, date of occurrence, property involved, participants, etc. For example, the data structure of key data about a property-related event is as follows:

[0043] {

[0044] "Event ID":"E12345",

[0045] "Event Type":"Delayed Delivery",

[0046] "Occurrence date":"2024-08-10",

[0047] "Real estate involved":{

[0048] "Property ID":"P9876",

[0049] "Address":"No. 123, City Center Road",

[0050] "Property Type":"Commercial"

[0051] },

[0052] "Participants":[

[0053] {"Participant ID":"C123",

[0054] "Name":"Zhang San",

[0055] "Role":"Home Buyer",

[0056] "Contact information":"123456789"},

[0057] {"Participant ID":"C456",

[0058] "Name":"Li Si",

[0059] "Role":"Developer",

[0060] "Contact information":"987654321"} ]

[0062] }.

[0063] Step 108: Perform statistical analysis on the key data using a statistical algorithm to obtain statistical analysis results.

[0064] In this step, the above structured key data must first be preprocessed to meet the statistical analysis requirements, including data type conversion and missing value processing. Then, statistical analysis is performed based on the preprocessed key data, such as calculating the number of occurrences of each type of event in different unit time, the number of participants, the average time for event resolution, etc., to obtain statistical analysis results. These statistical analysis results reveal important patterns in key data and trends that change over time and social dynamics, ensuring the depth and breadth of data analysis.

[0065] Step 110: Update the pre-constructed first analysis template according to the statistical analysis result to obtain a second analysis template; wherein the first analysis template is determined according to the historical report data.

[0066] Specifically, the first analysis template is determined based on the historical report data. The first analysis template contains fixed report content and chapters corresponding to the historical report data. Different types of historical report data correspond to different first analysis templates. For example, housing-related events and disaster-related events both have corresponding first analysis templates. If the statistical analysis results contain content that is not included in the first analysis template, the first analysis template needs to be updated according to the statistical analysis results to obtain a second analysis template so that the second analysis template can display trends and insights of data changes over time.

[0067] Step 112: Fill in the second analysis template based on the key data to generate the analysis report.

[0068] Specifically, the corresponding content for filling the second analysis template is determined from the key data. The filling process includes but is not limited to formatting the key data to adapt to the report requirements, generating visual elements, etc. The key data can be automatically filled into the reserved position of the template using a script or a report generation tool. The second analysis template includes a plurality of template elements, which are used to indicate the content that needs to be filled into the template. Each template element has a corresponding reserved position. For each template element in the second analysis template, the key data corresponding to the template element is filled into the reserved position corresponding to the template element in a predetermined form. Among them, the predetermined form includes visual charts or analytical texts, etc. For visual charts, a chart automatic generation tool can be used to automatically generate them according to the key data. For analytical texts, a large language model can be used to automatically generate descriptive and analytical texts, including background information, explanatory trends, and possible causes and suggestions. Finally, a complete analysis report is generated.

[0069] Through the method of this embodiment, the analysis report can not only reflect the current data status, but also show the trends and insights of data changes over time. The dynamic adjustment of the analysis template ensures that the analysis report always remains relevant and practical, supporting enterprises to make informed decisions in a changing market environment. The method of extracting key information according to entities, relationships, and associations in this embodiment expands the depth of data analysis, statistical analysis of key data expands the breadth of data analysis, and real-time updating and personalized display of the content of the analysis report greatly improves the strategic value and operational efficiency of the analysis report.

[0070] Based on the above steps 102 to 112, the data report intelligent generation method based on multidimensional feature mining provided by the present application includes obtaining clue data to be processed; determining the entities and attributes to be extracted in the clue data according to the pre-extracted core analysis elements, and establishing the association relationship between the entities; wherein the core analysis elements are determined according to the historical report data, so as to accurately determine the specific content to be extracted from the clue data. Construct a first prompt word according to the entity, the attribute, the association relationship and the clue data, input the first prompt word into the large language model, and output the key data corresponding to the clue data through the large language model; wherein the first prompt word is used to instruct the large language model to extract key data from the clue data according to the entity, the attribute and the association relationship. Utilize the understanding ability of the large language model, instruct the large language model to extract key data through the first prompt word, so as to provide a data basis for the subsequent generation of an analysis report. Perform statistical analysis on the key data using a statistical algorithm, mine new trends and new content in the key data, and obtain statistical analysis results. The first analysis template constructed in advance is updated according to the statistical analysis result to obtain the second analysis template; wherein the first analysis template is determined according to the historical report data; and the dynamic adjustment of the analysis template is realized to adapt to the real-time change trend of the clue data. The second analysis template is filled based on the key data to generate the analysis report. Automatically generating the analysis report helps to improve the generation rate and accuracy of the analysis report, ensure that the report content is updated as the data is updated, and meet the needs of the rapidly changing decision-making environment.

[0071] The innovation of this application lies in that it combines structured data extraction, statistical analysis with dynamic template adjustment, and data-driven report generation, providing an efficient, accurate and dynamic report generation solution for the political and legal fields. By using advanced large language models to extract key data from unstructured data and converting it into a structured format, the efficiency of data processing and the accuracy of analysis are significantly improved. In addition, through in-depth statistical analysis of the extracted key data, not only can new statistically significant patterns be identified, but also report templates can be dynamically adjusted based on these insights. This ensures that as time goes by and data characteristics change, the generated reports continue to reflect the latest trends and insights, thereby supporting more effective decision making. Through intelligent data processing and report generation, the operational efficiency and responsiveness of political and legal organs are enhanced.

[0072] In some embodiments, the method for determining the core analysis elements is as follows: Figure 2 , including the following steps:

[0073] Step 202: Obtain historical report data, and perform data cleaning on the historical report data;

[0074] Step 204: determining an initial analysis template based on the historical report data that has undergone data cleaning; wherein the initial analysis template includes a plurality of template elements;

[0075] Step 206: extract all template elements in the initial analysis template, summarize all template elements, and determine the core analysis elements.

[0076] Specifically, the historical report data is an analysis report of a historical period. Data cleaning of the historical report data is to ensure the accuracy and consistency of the data and provide a high-quality data foundation for subsequent data mining and analysis. In a specific embodiment, step 202 further includes: ① Removing unnecessary characters and interfering characters in the historical report data: automatically identifying and removing unnecessary characters and interfering characters in the text, such as extra spaces and special symbols, to remove characters that are irrelevant to the text content. ② Standardizing the punctuation marks in the historical report data: converting Chinese punctuation marks into English punctuation marks, or converting English punctuation marks into Chinese punctuation marks according to the language environment of the text to ensure the consistency of the text. At the same time, regular expressions can also be used to remove unnecessary repeated punctuation marks in the text and delete redundant punctuation marks. ③ Correcting spelling errors in the historical report data using natural language processing technology: In the data cleaning process, correcting spelling errors is a key link to ensure text quality and data accuracy. Use natural language processing technology to automatically detect spelling errors in the text, and assist in spelling correction through context analysis to ensure the use of contextually appropriate vocabulary, such as correcting the misspelled "set defense" to "survey room". Text error correction not only improves the readability of the text, but also enhances the accuracy of data analysis, laying a solid foundation for subsequent automated processing and report generation.

[0077] After the historical report data is cleaned, an initial analysis template is determined based on the cleaned historical report data. Step 204 specifically includes:

[0078] Step 2042: divide the cleaned historical report data into paragraphs to obtain multiple text paragraphs.

[0079] Specifically, each historical report is automatically divided into paragraphs, and the segmentation can be achieved by identifying specific formatting marks, such as line breaks, paragraph numbers or titles, etc. For example, if there are 100 historical report data, each report is segmented, and finally 810 text paragraphs are obtained.

[0080] Step 2044: cluster the multiple text paragraphs using a clustering algorithm to obtain multiple clusters.

[0081] Specifically, text embedding models or other vectorization techniques are used to convert text data into feature vectors in numerical form for machine learning processing. If the feature dimension of text paragraph data is very high, methods such as principal component analysis (PCA) can be applied to reduce the dimension to simplify subsequent cluster analysis. According to the nature of the data, a suitable clustering algorithm is selected, such as K-means clustering algorithm (K-means), hierarchical clustering or density-based spatial clustering of applications with noise (DBSCAN). The clustering algorithm is applied to the feature vector of the text paragraph, and the text paragraphs analyzing the same problem are grouped into the same cluster, each cluster representing the analysis of a class of problems or content patterns. The effect of clustering is evaluated according to indicators such as the silhouette coefficient, and the parameters of the clustering algorithm (such as the number of clusters) are adjusted to optimize the algorithm.

[0082] Step 2046: For each text paragraph in each cluster, the large language model is used to extract the core analysis information corresponding to the text paragraph in the cluster, including:

[0083] For each cluster, a plurality of text paragraphs are randomly selected from the cluster, a third prompt word is generated according to the plurality of text paragraphs, the third prompt word is input into the large language model, and the core analysis information is output through the large language model; wherein the third prompt word is used to instruct the large language model to summarize and analyze the plurality of text paragraphs to obtain the core analysis information.

[0084] Specifically, each cluster contains several text paragraphs, and multiple text paragraphs, such as three text paragraphs, are randomly selected from the multiple text paragraphs. By constructing a clear prompt (i.e., the third prompt word), a large language model is used to automatically learn and extract common core analysis information from multiple text paragraphs. For example, the third prompt word can be:

[0085] "

[0086] Please analyze the following text paragraphs from the same cluster and summarize their common core analysis points:

[0087] [Insert text paragraph 1]

[0088] [Insert text paragraph 2]

[0089] [Insert text paragraph 3]

[0090] "

[0091] The third prompt word is sent to the large language model for processing, and the content returned by the large language model is received, such as the output of the large language model: "clue type proportion analysis", to check whether the large language model accurately captures the core theme and focus of the text. If the check passes, the core analysis information is "clue type proportion analysis". If the check fails, the design of the third prompt word can be adjusted according to the return content of the large language model to optimize the output result of the large language model.

[0092] Step 2048: construct a second prompt word based on the core analysis information and the historical report data, input the second prompt word into the large language model, and output an initial analysis template corresponding to the historical report data through the large language model; wherein the second prompt word is used to instruct the large language model to replace the variable information in the historical report data with the corresponding template elements.

[0093] Specifically, the initial analysis template defines the basic framework and content requirements of the subsequent analysis report. The large language model is used to automatically identify and replace specific data and specific identifiers from the text paragraphs, and convert them into common template elements (or placeholders) to create a flexible and reusable initial analysis template. Specifically, the second prompt word is constructed to instruct the large language model to replace the variable information in the historical report data with the corresponding template elements.

[0094] Exemplarily, the second prompt word may be:

[0095] "

[0096] 'This year, we have collected and handled 2,000 intelligence clues, of which 1,200 were related to housing.'

[0097] Please replace any specific data and specific logos with appropriate placeholders to form a reusable template.

[0098] "

[0099] Among them, "2,000 intelligence clues have been collected and handled this year, including 1,200 clues related to housing." is part of the text paragraph in the historical report data. At the same time, core analysis information can also be added to the second prompt word to ensure that the large language model can accurately understand the core ideas of the text paragraph in the historical report data, which helps the large language model to identify information such as event types in the text paragraph.

[0100] The constructed second prompt word is sent to the large language model for processing. The large language model automatically replaces the specific variable information identified in the text paragraph with template elements such as [total number of events throughout the year] and [number of housing-related events throughout the year] based on the second prompt word. Continuing with the previous example, the initial analysis template output by the large language model is: "Intelligence clues collected and handled since the beginning of this year [total number of events throughout the year], including housing-related clues [number of housing-related events throughout the year]". When generating the analysis report later, the extracted "total number of events throughout the year" and "housing-related events throughout the year" values ​​can be directly filled into the corresponding positions of the template elements.

[0101] After the initial analysis template is determined, it is necessary to extract the core analysis elements from the initial analysis template. The core analysis elements are the key basis for clue data extraction and analysis report generation. For example, in the template of the housing-related incident analysis report, although the initial analysis template is filled with multiple data-related template elements such as "total number of incidents", "number of housing-related incidents" and "proportion of various types of incidents", it is found after analysis that "clue type" is the common support of these template elements, that is, the core analysis element. By correctly classifying each actual clue data into its "clue type", the template element information such as "number of housing-related incidents" and "proportion of incidents" can be automatically calculated through statistical analysis. Determining these core analysis elements is conducive to formulating clear guidance for subsequent data extraction and processing work, ensuring the accuracy of the analysis report content and the effectiveness of in-depth analysis.

[0102] Furthermore, all template elements in the initial analysis template are collected, such as [Total number of events in the whole year], [Number of events involving housing categories in the whole year], [Proportion of events involving housing categories], etc. The data type and content represented by each template element are analyzed, including the classification of data such as values, percentages, and names.

[0103] Exemplarily, the initial analysis template for security incidents is as follows:

[0104] "A total of [total number of incidents for the whole year] safety incidents were handled this year. Among them, there were [number of fire incidents] fire incidents, accounting for [percentage of fire incidents]%; there were [number of traffic accidents] traffic accidents, accounting for [percentage of traffic accidents]%. Among these incidents, the most common was [most common incident type], involving [number of most common incidents] times."

[0105] Extract all template elements in the above initial analysis template, including: [Total number of events throughout the year] [Number of fire events] [Proportion of fire events] [Number of traffic accidents] [Proportion of traffic accidents] [Most common event type] [Number of most common events involved].

[0106] When summarizing and extracting the core analysis elements, analyze the relationship between each template element and the analysis report theme. First, observe whether these template elements change around a specific theme or classification, such as "housing-related clues". Then, identify the dependencies between template elements. For example, if multiple template elements such as [Number of housing-related events throughout the year] and [Proportion of housing-related events] are dependent on the clue category "housing-related", this shows that "clue category" is a key element. For another example, in the aforementioned initial analysis template, fire events and traffic accidents are specifically mentioned, indicating that "type" is a key element, and each type has a corresponding number of events and proportion, showing the dependency between template elements. If multiple important template elements are associated with a specific category (such as a clue category), then this category is likely to be a core analysis element that needs special attention when building the template. Then, verify the summarized core analysis elements to ensure that it meets the actual data content and report requirements. The verification process can be confirmed with the data manager or business analyst.

[0107] In the above example, all important template elements are directly related to the event type, so the core analysis element is determined to be "event type" because the classification and analysis of data are based on different event types (such as fire, traffic accident).

[0108] After the core analysis elements are determined, the initial analysis template needs to be further updated to make the determined analysis template more flexible and practical, and to adapt to different types of data analysis requirements. In some embodiments, the method for determining the first analysis template includes: determining a newly added template element based on the core analysis elements; adding the newly added template element to the initial analysis template to obtain the first analysis template.

[0109] Specifically, the identified core analysis elements and their application in the initial analysis template are recorded in detail as a reference for future template updates and maintenance. Then the template is updated based on the core analysis elements to ensure that all related template elements can accurately reflect the changes in the core analysis elements. Continuing with the previous example, the core analysis element is "Event Type", and the newly added template element is [Event Type]. The newly added template element is added to the initial analysis template to obtain the first analysis template:

[0110] "A total of [total number of incidents for the whole year] security incidents were handled this year. Among them, [incident type 1] incidents [number of incident type 1] times, accounting for [percentage of incident type 1]%; [incident type 2] incidents [number of incident type 2] times, accounting for [percentage of incident type 2]%. Among these incidents, the most common one was [most common incident type], involving [number of most common incidents involved] times."

[0111] Through the above analysis template update process, not only the core analysis elements in the analysis template are identified, but also the design of the analysis template can be optimized and standardized according to the core analysis elements, making the analysis template more flexible and practical, and able to adapt to different types of data analysis needs.

[0112] In some embodiments, the first analysis template includes a plurality of template elements; and updating the pre-constructed first analysis template according to the statistical analysis result to obtain the second analysis template includes:

[0113] A newly added template element is determined according to the statistical analysis result, and the newly added template element is added to the first analysis template to obtain the second analysis template.

[0114] Specifically, in the aforementioned embodiments, the unstructured clue data has been processed by technical methods such as large language models, key data has been extracted, and structured storage has been performed. Since the clue data changes in real time, there may be a mismatch between the currently acquired clue data and the various template elements in the determined first analysis template. Therefore, it is also necessary to update the first analysis template according to the statistical analysis results of the key data, that is, to dynamically adjust the first analysis template, so as to ensure that the subsequent analysis report can accurately reflect the current data status and capture and display the missing changes that change over time and social dynamics. The statistical analysis results include new patterns or trends with statistical significance. For example, the new trend includes an increase in delayed delivery events in a specific season. A chapter specifically analyzing this phenomenon can be added to the analysis template as a newly added template element to explore the reasons and make suggestions. The newly added template element is added to the first analysis template, and the second analysis template is obtained after the first analysis template is updated. At the same time, the charts, data visualization elements and text descriptions in the analysis template can also be adaptively updated to ensure that the second analysis template can reflect the latest data and analysis results. Afterwards, each template element in the second analysis template is filled in according to the key data, and an analysis report is generated so that the analysis report can continuously reflect the latest trends and insights, thereby supporting more effective decision making.

[0115] It should be noted that the method of the embodiment of the present application can be performed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present application, and the multiple devices will interact with each other to complete the described method.

[0116] It should be noted that the above describes some embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the above embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0117] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides an intelligent generation device for analysis reports based on multi-dimensional feature mining.

[0118] refer to Figure 3 , the intelligent generation device of analysis report based on multi-dimensional feature mining comprises:

[0119] An acquisition module 302 is configured to acquire clue data to be processed;

[0120] The determination module 304 is configured to determine the entities and attributes to be extracted in the clue data according to the pre-extracted core analysis elements, and establish association relationships between the entities; wherein the core analysis elements are determined according to the historical report data;

[0121] The extraction module 306 is configured to construct a first prompt word according to the entity, the attribute, the association relationship and the clue data, input the first prompt word into the large language model, and output key data corresponding to the clue data through the large language model; wherein the first prompt word is used to instruct the large language model to extract key data from the clue data according to the entity, the attribute and the association relationship;

[0122] The analysis module 308 is configured to perform statistical analysis on the key data using a statistical algorithm to obtain statistical analysis results;

[0123] The updating module 310 is configured to update the pre-built first analysis template according to the statistical analysis result to obtain a second analysis template; wherein the first analysis template is determined according to the historical report data;

[0124] The generating module 312 is configured to fill the second analysis template based on the key data to generate the analysis report.

[0125] In some embodiments, the first analysis template includes multiple template elements; the update module 310 is further configured to determine newly added template elements according to the statistical analysis results, and add the newly added template elements to the first analysis template to obtain the second analysis template.

[0126] In some embodiments, the determination module 304 is further configured to obtain historical report data and perform data cleaning on the historical report data;

[0127] Determining an initial analysis template based on historical report data that has undergone data cleaning; wherein the initial analysis template includes a plurality of template elements;

[0128] All template elements in the initial analysis template are extracted, all template elements are summarized and generalized, and the core analysis elements are determined.

[0129] In some embodiments, the determination module 304 is further configured to divide the historical report data after data cleaning into paragraphs to obtain a plurality of text paragraphs;

[0130] Clustering algorithms are used to cluster multiple text paragraphs to obtain multiple clusters;

[0131] For each text paragraph in each cluster, the core analysis information corresponding to the text paragraph in the cluster is extracted using a large language model;

[0132] A second prompt word is constructed according to the core analysis information and the historical report data, the second prompt word is input into the large language model, and an initial analysis template corresponding to the historical report data is output through the large language model; wherein the second prompt word is used to instruct the large language model to replace the variable information in the historical report data with the corresponding template elements.

[0133] In some embodiments, the determination module 304 is further configured to remove unnecessary characters and interfering characters in the historical report data; standardize punctuation marks in the historical report data; and correct spelling errors in the historical report data using natural language processing technology.

[0134] In some embodiments, the determination module 304 is further configured to randomly select multiple text paragraphs from each cluster, generate a third prompt word based on the multiple text paragraphs, input the third prompt word into the large language model, and output the core analysis information through the large language model; wherein the third prompt word is used to instruct the large language model to summarize and analyze the multiple text paragraphs to obtain the core analysis information.

[0135] In some embodiments, the determination module 304 is further configured to determine a newly added template element based on the core analysis element; and add the newly added template element to the initial analysis template to obtain the first analysis template.

[0136] For the convenience of description, the above device is described in terms of functions divided into various modules. Of course, when implementing the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0137] The device of the above embodiment is used to implement the corresponding data report intelligent generation method based on multidimensional feature mining in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.

[0138] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments and methods, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for intelligently generating data reports based on multidimensional feature mining as described in any of the above embodiments is implemented.

[0139] Figure 4 A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.

[0140] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0141] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0142] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0143] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0144] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0145] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0146] The electronic device of the above-mentioned embodiment is used to implement the corresponding data report intelligent generation method based on multi-dimensional feature mining in any of the above-mentioned embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.

[0147] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the data report intelligent generation method based on multidimensional feature mining as described in any of the above embodiments.

[0148] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0149] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the data report intelligent generation method based on multidimensional feature mining as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0150] Based on the same concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a computer program product, including computer program instructions. When the computer program instructions are run on a computer, the computer executes the method described in any of the above embodiments, which has the beneficial effects of the corresponding method embodiments and will not be repeated here.

[0151] It is understandable that before using the technical solutions of each embodiment of the present disclosure, the type, scope of use, usage scenarios, etc. of the personal information involved will be informed to the user in an appropriate manner, and the user's authorization will be obtained.

[0152] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly remind the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can independently choose whether to provide personal information to software or hardware such as an electronic device, application, server or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0153] As an optional but non-limiting implementation, in response to receiving the user's active request, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to the electronic device providing personal information.

[0154] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0155] A person skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present application is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0156] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present application difficult to understand, the known power supply / ground connection with the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform to be implemented in the embodiments of the present application (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuits) to describe exemplary embodiments of the present application, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered to be illustrative rather than restrictive.

[0157] Although the present application has been described in conjunction with specific embodiments of the present application, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0158] The embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the present application. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the protection scope of the present application.

Claims

1. A method for intelligently generating data reports based on multidimensional feature mining, characterized in that: include: Get the clue data to be processed; Determine entities and attributes to be extracted in the clue data according to the pre-extracted core analysis elements, and establish association relationships between entities; wherein the core analysis elements are determined according to the historical report data; Construct a first prompt word according to the entity, the attribute, the association relationship and the clue data, input the first prompt word into a large language model, and output key data corresponding to the clue data through the large language model; wherein the first prompt word is used to instruct the large language model to extract key data from the clue data according to the entity, the attribute and the association relationship; Performing statistical analysis on the key data using a statistical algorithm to obtain statistical analysis results; According to the statistical analysis result, the pre-constructed first analysis template is updated to obtain a second analysis template; wherein the first analysis template is determined according to the historical report data; The second analysis template is filled based on the key data to generate the analysis report.

2. The method according to claim 1, characterized in that The first analysis template includes a plurality of template elements; the pre-constructed first analysis template is updated according to the statistical analysis result to obtain a second analysis template, including: A newly added template element is determined according to the statistical analysis result, and the newly added template element is added to the first analysis template to obtain the second analysis template.

3. The method according to claim 1, characterized in that The method for determining the core analysis elements includes: Acquire historical report data, and perform data cleaning on the historical report data; Determining an initial analysis template based on historical report data that has undergone data cleaning; wherein the initial analysis template includes a plurality of template elements; All template elements in the initial analysis template are extracted, all template elements are summarized and generalized, and the core analysis elements are determined.

4. The method according to claim 3, characterized in that Determining the initial analysis template based on the historical report data after data cleaning includes: Divide the cleaned historical report data into paragraphs to obtain multiple text paragraphs; Clustering algorithms are used to cluster multiple text paragraphs to obtain multiple clusters; For each text paragraph in each cluster, the core analysis information corresponding to the text paragraph in the cluster is extracted using a large language model; A second prompt word is constructed according to the core analysis information and the historical report data, the second prompt word is input into the large language model, and an initial analysis template corresponding to the historical report data is output through the large language model; wherein the second prompt word is used to instruct the large language model to replace the variable information in the historical report data with the corresponding template elements.

5. The method according to claim 3, characterized in that: Performing data cleaning on the historical report data, including: removing unnecessary characters and interfering characters from the historical report data; Standardizing punctuation marks in the historical report data; Natural language processing techniques are used to correct spelling errors in the historical reporting data.

6. The method according to claim 4, characterized in that For each text paragraph in each cluster, the core analysis information corresponding to the text paragraph in the cluster is extracted using a large language model, including: For each cluster, a plurality of text paragraphs are randomly selected from the cluster, a third prompt word is generated according to the plurality of text paragraphs, the third prompt word is input into the large language model, and the core analysis information is output through the large language model; wherein the third prompt word is used to instruct the large language model to summarize and analyze the plurality of text paragraphs to obtain the core analysis information.

7. The method according to claim 3, characterized in that The method for determining the first analysis template comprises: Determining new template elements based on the core analysis elements; The newly added template elements are added to the initial analysis template to obtain the first analysis template.

8. A data report intelligent generation device based on multi-dimensional feature mining, characterized in that: include: An acquisition module, configured to acquire clue data to be processed; A determination module, configured to determine entities and attributes to be extracted in the clue data according to the pre-extracted core analysis elements, and to establish association relationships between entities; wherein the core analysis elements are determined according to the historical report data; an extraction module, configured to construct a first prompt word according to the entity, the attribute, the association relationship and the clue data, input the first prompt word into a large language model, and output key data corresponding to the clue data through the large language model; wherein the first prompt word is used to instruct the large language model to extract key data from the clue data according to the entity, the attribute and the association relationship; an analysis module, configured to perform statistical analysis on the key data using the large language model to obtain statistical analysis results; An updating module, configured to update a pre-built first analysis template according to the statistical analysis result to obtain a second analysis template; wherein the first analysis template is determined according to historical report data; A generation module is configured to fill in the second analysis template based on the key data to generate the analysis report.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to enable a computer to execute the method according to any one of claims 1 to 7.

Images