Hierarchical table data visualization intelligent construction method and device based on natural language questions and answers, computer readable storage medium and electronic equipment

Abstract

The invention relates to a hierarchical table data visualization intelligent construction method and device based on natural language questions and answers. The method comprises the steps that a user uploads hierarchical table data in an Excel format; a user puts forward a visualization requirement on the hierarchical table by adopting a natural language interaction mode; the system analyzes the natural language to determine a data area and a visualization task; the system extracts auxiliary hierarchical table transformation according to a data screening result and data insight; and the system defines the row / column descriptor priority according to the table unit specified by the user so as to interact and realize visualization. According to the method, through a natural language processing technology, a user can propose a data analysis demand in a natural language mode; through automatic data insight identification, a user is helped to discover valuable information in the data; through an intelligent data visualization construction process, a visual result satisfying a user is rapidly generated. According to the invention, the threshold of data analysis is reduced, and the efficiency and accuracy of data analysis are improved.

Description

Technical Field

[0001] The present application belongs to the field of data visualization technology, and in particular relates to a method, device, computer-readable storage medium, and electronic device for intelligently constructing hierarchical table data visualization based on natural language question and answer. Background Art

[0002] With the rapid development of information technology, the amount of data has shown explosive growth, and data analysis and processing have become key requirements in various fields. Among the many data expression forms, hierarchical table data is widely used in scientific research, finance, business intelligence, medical health and other fields due to its powerful information organization ability. Hierarchical table data can express complex data relationships and multi-dimensional data features through multi-level headers of rows and columns, making data storage and query more efficient. However, existing hierarchical table data visualization technologies face a series of challenges. First, traditional visualization technologies usually require users to have a certain technical background to operate complex software interfaces to filter, process and visualize data, which limits the use of non-professional users and increases the threshold for data analysis; second, existing visualization technologies often lack the ability to understand and process natural language, and users cannot directly use natural language to describe their analysis needs, resulting in users spending extra effort to learn how to convert natural language requirements into understandable operating instructions for visualization tools. In addition, the complexity of hierarchical table data also brings challenges to visualization. The data in hierarchical tables are often multidimensional and nested. Existing visualization technologies are difficult to effectively process such complex data structures and difficult to intuitively display hierarchical relationships and related information between data. Finally, existing visualization technologies often encounter performance bottlenecks when processing large-scale data, resulting in a slow or even incomplete visualization process. In order to solve the above problems, researchers have tried to combine natural language processing technology with data visualization technology to provide a more intuitive and efficient means of data analysis. Through the natural language interface, users can put forward query requirements in natural language, and the system automatically understands user requirements and generates corresponding visualization results. However, although this research direction has great potential, it still faces many challenges.

[0003] First, the ambiguity and diversity of natural language make it difficult for the system to accurately understand the user's query intent. For example, users may ask the same data in many different ways, such as "show sales trends" or "sales changes over time", and the system needs to be able to recognize the same needs behind these different expressions. Second, most existing natural language visualization systems focus on simple data sets and lack effective support for complex, multidimensional, and hierarchical tabular data. The complexity of hierarchical tabular data lies in the nested relationships and multi-level headers between its data, which requires the system to be able to deeply understand and analyze the data. In addition, automated data insight recognition is another challenge. When analyzing data, users often hope to discover patterns, trends, and anomalies in the data, which requires the system to have certain data insights. However, most existing systems lack this ability, and users still need to discover this information through manual exploration. Finally, existing natural language visualization systems lack flexibility in the visualization construction process and have a low degree of automation. After obtaining the visualization results, users often need to make multiple adjustments and optimizations, which is cumbersome and inefficient. Summary of the invention

[0004] In view of the above-mentioned defects in the prior art, the present invention proposes a new intelligent construction method for hierarchical tabular data visualization based on natural language question answering (natural language for hierarchical tabular data visualization, hereinafter referred to as NL4HiTV system). This method uses natural language processing technology to enable users to propose data analysis requirements in natural language; through automated data insight recognition, it helps users discover valuable information in the data; through intelligent data visualization construction process, it quickly generates visualization results that satisfy users. This application aims to lower the threshold for data analysis, improve the efficiency and accuracy of data analysis, and help users better understand and use hierarchical tabular data.

[0005] In order to achieve the above objectives, this application provides the following technical solutions:

[0006] The first aspect of the present application provides a method for intelligently constructing hierarchical table data visualization based on natural language question answering, the method comprising:

[0007] Users upload hierarchical table data in Excel format;

[0008] Users use natural language interaction to express their needs for visualization of hierarchical tables;

[0009] The system parses natural language to determine data areas and visualization tasks;

[0010] The system extracts auxiliary hierarchical table transformations based on the results of data screening and data insights;

[0011] The system interactively implements visualization based on user-specified table cell definition row / column descriptor priorities.

[0012] Optionally, in the method of the present application, the user uploads hierarchical table data in Excel format, including:

[0013] The user uploads hierarchical table data in Excel format, and the system automatically parses the hierarchical table data to generate table data in JSON format;

[0014] Convert the table data in JSON format into the input format required by the question-answering model for model input and front-end table display;

[0015] The system extracts the header structure of the hierarchical table and integrates all the data attributes in the table to obtain the name, type and numerical result of each cell.

[0016] Optionally, in the method of the present application, the user proposes a visualization requirement for the hierarchical table in an interactive manner using natural language, including:

[0017] The system provides users with a natural language interaction interface, supporting users to propose visualization requirements through natural language interaction;

[0018] Users use natural language interaction to express their needs for visualization of hierarchical tables;

[0019] The system uses a question-answering model and a rule-based approach to analyze the natural language proposed by the user and convert the natural language into corresponding operations of the visualization system.

[0020] Optionally, in the method of the present application, the system parses natural language to determine data areas and visualization tasks, including:

[0021] When users enter queries in natural language, the system first parses these queries into JSON formatted data;

[0022] The system converts the output of the question-answering model into a filter expression, which is organized in a list format, with each list element representing a filter condition corresponding to a specific data subset in the hierarchical table;

[0023] The system uses the Word2Vec model and similarity matching algorithm to optimize and adjust the filter list;

[0024] The system generates a syntactic analysis tree through a syntactic analysis tool. The tree structure of the syntactic analysis tree represents the grammatical structure of the sentence. The system identifies the keywords and phrases in the user query sentence based on the syntactic analysis tree, and then determines the data area and visualization task through syntactic analysis.

[0025] The system uses keyword recognition methods to map natural language query statements to visualization tasks supported by the system.

[0026] Optionally, in the method of the present application, the Word2Vec model converts words in the text into vectors, and identifies semantically similar words by calculating the similarity between the vectors. The system uses the Word2Vec model to compare the user's natural language query with the table header in the table to determine the most relevant data area;

[0027] The visualization tasks supported by the system include: bar charts, line charts, pie charts, scatter plots, box plots, heat maps, tree maps, ring charts, radar charts, dot plots, stacked charts, area charts and word clouds.

[0028] Optionally, in the method of the present application, the system extracts auxiliary hierarchical table transformation according to the results of data screening and data insight, including:

[0029] The system reorganizes the data through multi-dimensional transformation methods and compares the data performance under different dimensional combinations;

[0030] The system uses box plots and standard deviation analysis to identify outliers, helping users discover potential data errors or important special events;

[0031] The system quantifies the linear or nonlinear relationship between different data dimensions by calculating the Pearson or Spearman correlation coefficient;

[0032] The system uses regression analysis to analyze data trends, identify linear trends, polynomial trends or seasonal patterns in the data, and provide users with a basis for future predictions;

[0033] The system uses statistical functions to quickly locate the maximum and minimum values ​​in a data subset, helping users identify key performance indicators or abnormal situations;

[0034] The system uses a quantitative evaluation mechanism to assess the contribution of different transformation methods to data insights;

[0035] The system has a feedback mechanism, and every evaluation and adjustment made by the user is used as data input to optimize the system's analysis algorithm and achieve self-learning and continuous improvement.

[0036] Optionally, in the method of the present application, the system defines the row / column descriptor priority according to the table cell specified by the user and then interactively realizes visualization, including:

[0037] When users input query statements in natural language, the system first parses these statements to determine the user's visualization requirements, then automatically selects a matching visualization library to generate the corresponding chart and embeds it in the appropriate location in the user interface;

[0038] The system provides an interactive visualization editing function. Users click on the embedded visualization results and select different visualization types or adjust the current visualization parameters through the pop-up editing panel.

[0039] Based on the similarity of data structure and the user's historical interaction behavior, the system recommends similar areas using the same visualization form by identifying patterns in the data;

[0040] The system recommends visualizations based on the dimensions and type of data.

[0041] A second aspect of the present application provides a device for intelligently constructing hierarchical table data visualization based on natural language question answering, the device comprising:

[0042] Data upload and processing module: used to upload hierarchical table data in Excel format, and parse, extract and process the data;

[0043] Request receiving module: used to receive the visualization requirements for hierarchical tables proposed by users in a natural language interactive manner;

[0044] Task confirmation module: used to parse natural language to determine data areas and visualization tasks;

[0045] Task transformation module: used to extract auxiliary hierarchical table transformation based on data screening results and data insights;

[0046] Task execution module: used to define row / column descriptor priorities based on user-specified table cells and interactively implement visualization.

[0047] When running, the device implements the steps of the aforementioned method for intelligently constructing hierarchical table data visualization based on natural language question and answer.

[0048] A third aspect of the present application provides an electronic device, comprising: a memory and a processor;

[0049] Memory: used to store computer programs;

[0050] Processor: used to execute the computer program to implement the steps of the aforementioned method for intelligently constructing hierarchical table data visualization based on natural language question and answer.

[0051] The fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the aforementioned method for intelligently constructing hierarchical table data visualization based on natural language question and answer are implemented.

[0052] In summary, this application proposes a new method for intelligently constructing hierarchical table data visualization based on natural language question answering, which has the following advantages:

[0053] (1) This method realizes the intelligent visualization construction of hierarchical table data through natural language question answering, which reduces the requirements for users' technical background and solves the problem that non-professional users find it difficult to operate complex software interfaces for data visualization. At the same time, it improves the efficiency of data processing and visualization and ensures the accuracy and usability of the analysis results.

[0054] (2) This method supports modeling of complex tabular data with a hierarchical structure, and the model can serve as the basis for data transformation and visualization range recommendation. By transforming tabular data, users can easily explore different aspects of the same data. The rich visualization results created on this basis can help users understand complex tabular data. In the process of creating visualizations, the flexible table visualization unit declaration and recommendation mechanism eliminate the need for users to repeatedly select data areas and specify visualization parameters, which improves exploration efficiency and reduces operational burden.

[0055] (3) This method can automatically parse hierarchical table data in Excel format and convert it into an easy-to-process JSON format. This process not only provides a basis for subsequent data processing and visualization, but also achieves seamless connection between user needs and data operations through natural language processing technology. Users can put forward their needs in natural language, and the system can accurately understand and convert them into corresponding data operation instructions, which greatly simplifies the user's operation process.

[0056] (4) In terms of dynamic data visualization, this method can intelligently recommend and generate suitable visualization charts based on user needs and data characteristics. This dynamic recommendation mechanism, combined with the interactive editing function, enables users to easily adjust the parameters of the visualization chart, such as color, axis labels, etc., to meet personalized display needs. In addition, this method also provides intelligent recommendation and flexibility for data visualization. Users can freely define the priority of row / column descriptors, and the system interactively realizes visualization based on this.

[0057] (5) The intelligent recommendation and reusability of this method enables the visualization settings of specific data to be recorded and applied to other data with similar structures, thereby improving work efficiency. The system learns and optimizes itself through user feedback, continuously improving the accuracy of data processing and visualization, and ensuring the reliability of analysis results.

[0058] (6) This method helps users identify data patterns and trends more quickly through intuitive visualization results, thereby supporting more effective decision making. This method lowers the technical threshold for using data visualization tools, and even non-technical users can easily use this method, enhancing data insight and helping users discover the deep meaning and potential value in the data.

[0059] Other features and advantages of the present application will be described in the following description, and partly become apparent from the description, or understood by practicing the present application. The purpose and other advantages of the present application can be realized and obtained by the techniques indicated in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions of the embodiments of the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0061] Figure 1 An overall implementation flowchart of a method for intelligently constructing hierarchical table data visualization based on natural language question and answer provided in one embodiment of the present application.

[0062] Figure 2 The figure is a schematic diagram of the implementation process from original complex table data to visual results according to the embodiment method of the present application.

[0063] Figure 3 A diagram of the hierarchical table area selection process in the method of an embodiment of the present application.

[0064] Figure 4 This is the overall architecture diagram of the hierarchical table question-and-answer model in the embodiment method of this application.

[0065] Figure 5 Flow chart of the table transformation algorithm in the embodiment of the present application.

[0066] Figure 6This is the visual analysis interface of the NL4HiTV system of the embodiment of the present application; in the figure, a represents the table transformation operation part; b represents the natural language query part; c represents the visualization parameter configuration part; d represents the table visualization part; e represents the visualization template recommended by the system for the table area.

[0067] Figure 7 A structural diagram of the composition of a device for intelligently constructing hierarchical table data visualization based on natural language question and answer provided in one embodiment of the present application.

[0068] Figure 8 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0069] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0070] As used herein, the term “including” and its variations are open inclusions, ie, “including but not limited to”; the term “based on” means “based at least in part on”; and the term “one embodiment” means “at least one embodiment”.

[0071] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless the context clearly indicates otherwise, it should be understood as "one or more".

[0072] Figure 1 The figure shows the overall implementation process of a method for intelligently constructing hierarchical table data visualization based on natural language question answering proposed in this application. This method belongs to the interactive data transformation and visualization construction technology for complex tables. Specifically, this method includes the following steps:

[0073] Users upload hierarchical table data in Excel format;

[0074] Users use natural language interaction to express their needs for visualization of hierarchical tables;

[0075] The system parses natural language to determine data areas and visualization tasks;

[0076] The system extracts auxiliary hierarchical table transformations based on the results of data screening and data insights;

[0077] The system interactively implements visualization based on user-specified table cell definition row / column descriptor priorities.

[0078] Figure 2 FIG. 1 shows the implementation process of converting original complex table data into visual results according to the method of the embodiment of the present application. Figure 2 As shown, the implementation process of the present application method includes the following five steps:

[0079] (1) Users upload hierarchical table data in Excel format

[0080] When users upload hierarchical table data in Excel format, the NL4HiTV system will automatically parse the table, generate table data in JSON format, and convert the table data into the input format required by the question-answering model for model input and front-end table display. At the same time, the system will extract the header structure of the hierarchical table and integrate all data attributes in the table to obtain the name, type and numerical result of each cell.

[0081] (2) Users use natural language interaction to express their needs for visualization of hierarchical tables

[0082] The NL4HiTV system provides users with a natural language interactive interface, supports users to propose visualization requirements through natural language interactive methods, and helps users intelligently build visualization results. The system uses a question-answering model and a rule-based method to analyze the natural language proposed by users, thereby converting natural language into corresponding operations of the visualization system.

[0083] (3) The system parses natural language to determine data areas and visualization tasks

[0084] In the NL4HiTV system, when users enter queries through natural language, the system first parses these queries into JSON formatted data. This step is critical because it allows the system to convert unstructured natural language into a structured format for subsequent processing. The system then converts the output of the question-answering model into filter expressions, which are organized in the form of lists, and each list element represents a filter condition corresponding to a specific subset of data in the hierarchical table. For example, if the user asks "sales for each quarter in 2020", the system will generate a filter expression to extract sales data for each quarter of 2020 from the table.

[0085] In order to improve the accuracy and relevance of the screening, the system uses the Word2Vec model and similarity matching algorithm to optimize the screening list. The Word2Vec model can convert words in the text into vectors and identify semantically similar words by calculating the similarity between vectors. The system uses this feature to compare the user's natural language query with the table headers in the table to determine the most relevant data area. The similarity matching algorithm further refines the screening conditions to ensure that the system can accurately identify and extract the data subsets that the user is interested in. After the optimization of the screening list is completed, the system generates a syntactic analysis tree through a syntactic analysis tool. This tree structure can represent the grammatical structure of the sentence, including phrase structure, dependency relationships, etc. The system analyzes these structures and identifies keywords and phrases, which are essential for understanding the user's query intent.

[0086] Finally, the system uses keyword recognition to map natural language query statements to one of the 13 visualization tasks supported by the system. For example, if the user's query contains trends that compare data from different time periods, the system maps this query to a trend analysis task and selects an appropriate visualization form (such as a line chart) to display the data. This mapping process ensures that the user's query intent can be accurately understood and executed.

[0087] (4) The system extracts auxiliary hierarchical table transformation based on data screening results and data insights

[0088] In the process of extracting auxiliary hierarchical table transformation based on the results of data screening and data insights, the system first reorganizes the data through multidimensional transformation methods to explore the data performance under different dimensional combinations, such as cross-analysis of time series and regional categories. In terms of outlier detection, the system uses statistical methods such as box plots and standard deviation analysis to identify values ​​that are significantly different from most data, thereby helping users discover potential data errors or important special events.

[0089] For correlation analysis between data, the system calculates the Pearson or Spearman correlation coefficient to quantify the linear or nonlinear relationship between different data dimensions. For example, the correlation between sales and advertising expenditures is analyzed to reveal the potential impact of market activities. Trend analysis is achieved through regression analysis methods. The system can identify linear trends, polynomial trends or seasonal patterns in the data, providing users with a basis for future predictions.

[0090] In extreme value analysis, the system uses statistical functions to quickly locate the maximum and minimum values ​​in a data subset, which helps users identify key performance indicators or anomalies. In order to evaluate the contribution of different transformation methods to data insights, the system uses a quantitative evaluation mechanism that can train models based on machine learning algorithms, historical data and user feedback, and score data transformation methods.

[0091] In addition, the system provides a wealth of user interaction functions, including drag and drop, selection and conversion operations, allowing users to manually adjust the data hierarchy and layout. Through intuitive data visualization tools such as bar charts, line charts, scatter plots and box plots, users can more intuitively understand the results of data transformation and verify the accuracy of system analysis.

[0092] The system also has a feedback mechanism designed. Every evaluation and adjustment made by the user is used as data input to optimize the system's analysis algorithm, achieve self-learning and continuous improvement. This design concept that combines automated analysis and user feedback not only improves the efficiency of data processing, but also ensures the accuracy and availability of analysis results, greatly enhancing the user's control and understanding of the data analysis process.

[0093] (5) The system defines the row / column descriptor priority based on the table cells specified by the user and then interactively implements visualization

[0094] In order to meet the diverse needs of users for visualization during data analysis, the system supports 13 different visualization forms, including bar charts, line charts, pie charts, scatter plots, box plots, heat maps, tree maps, ring charts, radar charts, dot charts, stacked charts, area charts, and word clouds. These visualization forms cover a variety of data analysis tasks, from basic comparison and distribution analysis to complex pattern recognition.

[0095] When users enter query statements through natural language, the system first parses these statements to determine the user's visualization needs. For example, if the user's query is "show the sales trend of each product category", the system will recognize that this is a time series analysis task and recommend using a line chart to display the data. The system will automatically select a suitable visualization library (such as D3.js or Echarts) to generate the corresponding chart and embed it in the appropriate location in the user interface.

[0096] To improve the user experience, the system provides interactive visualization editing capabilities. Users can click on the embedded visualization results and select different visualization types or adjust the parameters of the current visualization through the pop-up editing panel. For example, users can change the color of the bar chart, adjust the axis labels, modify the display of the data series, or switch to stacking mode to see the contribution of each part to the total. These adjustments can be made through simple sliders, drop-down menus, or color selectors, allowing users to easily customize the visualization display without professional knowledge.

[0097] In addition, the system also has an intelligent recommendation function that can identify patterns in the data and recommend visualization forms. For example, if a user is viewing sales data for a specific region and selects a bar chart to display it, the system will automatically analyze other similar data areas in the table (such as sales data for other regions) and recommend the same bar chart visualization form. This recommendation mechanism is based on the similarity of data structure and the user's historical interaction behavior, and aims to enhance comparison and analysis between data through a consistent visualization method.

[0098] When recommending visualizations, the system also considers the dimensions and types of data. For categorical data, the system may recommend a pie chart or bar chart to show proportions and distribution; for numerical data, it may recommend a line chart or area chart to show trends and changes. For data with multiple dimensions, the system may recommend a scatter plot or heat map to show the relationship between variables.

[0099] In terms of technical implementation, the system backend will build a mapping model between visualization tasks and visualization forms. This model is based on the natural language processing results of user input and combines data content and structural features to select the most appropriate visualization method. The front end is responsible for dynamically rendering the selected visualization form to the user interface and providing rich interactive controls to meet the user's personalized needs.

[0100] Figure 3 The hierarchical table area selection process in the embodiment of the present application is shown. The area selection process includes two functions (filter_tree h) and (filter_level l). Among them, h represents the header of the hierarchical table, and l represents the level of the hierarchical table. Functions should be called in sequence, and subsequent functions filter the return area of ​​the previous function. The function (filter_tree h) selects the corresponding subtree in the tree hierarchy according to the header cell. If h is a leaf header, it directly filters to the row or column of the corresponding area, otherwise it filters to the corresponding area of ​​the header index. The function (filter_level l) selects a subtree from the input structure according to the level in the hierarchical table, and returns the sub-area of ​​the header index on that level.

[0101] Figure 4 The overall architecture of the hierarchical table question-answering model in the embodiment of the present application is shown. Throughout the process, the parser of the table question-answering model needs to use BERT to encode the natural language questions and the headers of the hierarchical table data, and then input the encoded results into the parsing framework for processing. The parsing framework combines the question and the header data to generate an expression in logical form. Finally, the executor in the form of a hierarchical structure uses logical expressions to query and operate the table to generate the final answer.

[0102] Figure 5The flow of the table transformation algorithm in the embodiment of the present application is shown. After the data transformation module receives the data area from the natural language query interpretation module, it first needs to determine whether the data is a cell area or a data block area, so as to determine which algorithm to use to determine the table transformation method.

[0103] Figure 6 The visual analysis interface of the NL4HiTV system in the embodiment of the present application is shown. In the figure, a represents the table transformation operation part; b represents the natural language query part; c represents the visualization parameter configuration part; d represents the table visualization part; and e represents the visualization template recommended by the system for the table area.

[0104] Furthermore, based on the above technical solution of this application, the following technical expansions can also be performed:

[0105] (1) Further enhance natural language processing capabilities: By integrating more advanced natural language processing technologies, such as context-aware question-answering systems, the system’s ability to understand user query intent can be improved, thereby providing more accurate data visualization results.

[0106] (2) Multimodal data fusion: In addition to text data, multiple data types such as images and sounds can also be fused to provide richer data visualization forms.

[0107] (3) Enhance cross-platform compatibility: Further develop cross-platform data visualization tools so that they can run on different operating systems and devices to meet the needs of different users.

[0108] Figure 7 The present application shows a device for intelligently constructing hierarchical table data visualization based on natural language question answering, comprising:

[0109] Data upload and processing module: used to upload hierarchical table data in Excel format, and parse, extract and process the data;

[0110] Request receiving module: used to receive the visualization requirements for hierarchical tables proposed by users in a natural language interactive manner;

[0111] Task confirmation module: used to parse natural language to determine data areas and visualization tasks;

[0112] Task transformation module: used to extract auxiliary hierarchical table transformation based on data screening results and data insights;

[0113] Task execution module: used to define row / column descriptor priorities based on user-specified table cells and interactively implement visualization.

[0114] When the above device is running, the steps of the method for intelligently constructing hierarchical table data visualization based on natural language question and answer disclosed in this application are implemented.

[0115] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the device, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be realized by a special hardware-based system that performs the specified function or operation, or can be realized by a combination of special hardware and computer instructions.

[0116] like Figure 8 As shown, the embodiment of the present application also discloses an electronic device, including: a processor 310, a communication interface 320, a memory 330 for storing a computer program executable by the processor, and a communication bus 340. The processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 runs the executable computer program to implement the steps of the above-mentioned natural language question-answering based hierarchical table data visualization intelligent construction method.

[0117] It is understandable that, in addition to the memory and the processor, the electronic device may also include an input device such as a keyboard, an output device such as a display, and other communication modules. The input device, the output device, and other communication modules communicate with the processor via an I / O interface (i.e., an input / output interface).

[0118] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0119] Furthermore, the present application also discloses a computer-readable storage medium. When the instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device can execute the various steps of the method for intelligently constructing hierarchical tabular data visualization based on natural language question and answer disclosed in the present application.

[0120] In the context of the present application, a computer-readable storage medium may be a tangible medium, more specific examples of which would include a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0121] In particular, according to an embodiment of the present application, the process described in the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method for intelligently constructing hierarchical table data visualization based on natural language question answering disclosed in the present application. When the computer program is executed by a processing device, the above functions defined in the method of the embodiment of the present application are executed.

[0122] Although several specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present application. The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solutions formed by a specific combination of the above technical features, but also should cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above public concept.

[0123] Those skilled in the art should also understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for intelligently constructing hierarchical table data visualization based on natural language question answering, characterized in that: The method comprises: Users upload hierarchical table data in Excel format; Users use natural language interaction to express their needs for visualization of hierarchical tables; The system parses natural language to determine data areas and visualization tasks; The system extracts auxiliary hierarchical table transformations based on the results of data screening and data insights; The system interactively implements visualization based on user-specified table cell definition row / column descriptor priorities.

2. The method according to claim 1, characterized in that The user uploads hierarchical table data in Excel format, including: The user uploads hierarchical table data in Excel format, and the system automatically parses the hierarchical table data to generate table data in JSON format; Convert the table data in JSON format into the input format required by the question-answering model for model input and front-end table display; The system extracts the header structure of the hierarchical table and integrates all the data attributes in the table to obtain the name, type and numerical result of each cell.

3. The method according to claim 1, characterized in that The user proposes a visualization requirement for a hierarchical table by using a natural language interaction method, including: The system provides users with a natural language interaction interface, supporting users to propose visualization requirements through natural language interaction; Users use natural language interaction to express their needs for visualization of hierarchical tables; The system uses a question-answering model and a rule-based approach to analyze the natural language proposed by the user and convert the natural language into corresponding operations of the visualization system.

4. The method according to claim 1, characterized in that The system parses natural language to determine data areas and visualization tasks, including: When users enter queries in natural language, the system first parses these queries into JSON formatted data; The system converts the output of the question-answering model into a filter expression, which is organized in a list format, with each list element representing a filter condition corresponding to a specific data subset in the hierarchical table; The system uses the Word2Vec model and similarity matching algorithm to optimize and adjust the filter list; The system generates a syntactic analysis tree through a syntactic analysis tool. The tree structure of the syntactic analysis tree represents the grammatical structure of the sentence. The system identifies the keywords and phrases in the user query sentence based on the syntactic analysis tree, and then determines the data area and visualization task through syntactic analysis. The system uses keyword recognition methods to map natural language query statements to visualization tasks supported by the system.

5. The method according to claim 4, characterized in that The Word2Vec model converts words in a text into vectors and identifies semantically similar words by calculating the similarity between vectors. The system uses the Word2Vec model to compare the user's natural language query with the table headers in the table to determine the most relevant data area; The visualization tasks supported by the system include: bar charts, line charts, pie charts, scatter plots, box plots, heat maps, tree maps, ring charts, radar charts, dot plots, stacked charts, area charts and word clouds.

6. The method according to claim 1, characterized in that The system extracts auxiliary hierarchical table transformations based on the results of data screening and data insights, including: The system reorganizes the data through multi-dimensional transformation methods and compares the data performance under different dimensional combinations; The system uses box plots and standard deviation analysis to identify outliers, helping users discover potential data errors or important special events; The system quantifies the linear or nonlinear relationship between different data dimensions by calculating the Pearson or Spearman correlation coefficient; The system uses regression analysis to analyze data trends, identify linear trends, polynomial trends or seasonal patterns in the data, and provide users with a basis for future predictions; The system uses statistical functions to quickly locate the maximum and minimum values ​​in a data subset, helping users identify key performance indicators or abnormal situations; The system uses a quantitative evaluation mechanism to assess the contribution of different transformation methods to data insights; The system has a feedback mechanism, and every evaluation and adjustment made by the user is used as data input to optimize the system's analysis algorithm and achieve self-learning and continuous improvement.

7. The method according to claim 1, characterized in that The system interactively implements visualization based on the row / column descriptor priority defined by the user-specified table cell, including: When users input query statements in natural language, the system first parses these statements to determine the user's visualization requirements, then automatically selects a matching visualization library to generate the corresponding chart and embeds it in the appropriate location in the user interface; The system provides an interactive visualization editing function. Users click on the embedded visualization results and select different visualization types or adjust the current visualization parameters through the pop-up editing panel. Based on the similarity of data structure and the user's historical interaction behavior, the system recommends similar areas using the same visualization form by identifying patterns in the data; The system recommends visualizations based on the dimensions and type of data.

8. A device for intelligently constructing hierarchical table data visualization based on natural language question answering, characterized in that: The device comprises: Data upload and processing module: used to upload hierarchical table data in Excel format, and parse, extract and process the data; Request receiving module: used to receive the visualization requirements for hierarchical tables proposed by users in a natural language interactive manner; Task confirmation module: used to parse natural language to determine data areas and visualization tasks; Task transformation module: used to extract auxiliary hierarchical table transformation based on data screening results and data insights; Task execution module: used to define row / column descriptor priorities based on user-specified table cells and interactively implement visualization.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for intelligently constructing hierarchical table data visualization based on natural language question and answer as described in any one of claims 1 to 8 are implemented.

10. An electronic device, characterized in that: include: Memory and processor; Memory: used to store computer programs; Processor: used to execute the computer program to implement the steps of the method for intelligently constructing hierarchical table data visualization based on natural language question and answer as described in any one of claims 1-8.

Images