Data analysis method and device, electronic equipment and computer program product

Abstract

The invention provides a data analysis method and device, electronic equipment and a computer program product, and the method comprises the steps: carrying out the interpretation of answer data determined according to question query data through a pre-trained data interpretation model, and outputting the interpretation information of the answer data. The data interpretation model is obtained by inputting knowledge of a specific industry into an industry vertical large model of the specific industry to enable the industry vertical large model to learn the knowledge of the specific industry; the knowledge of the specific industry comprises knowledge fields, and further comprises at least one of field meta-knowledge, field default knowledge and field sequence knowledge. Through the arrangement, the data interpretation model obtained by knowledge learning based on the industry vertical large model can fully understand the data structure of the industry in which the question query data is located, so that the answer data is correctly and comprehensively interpreted, and the readability of the answer data is improved.

Description

Technical Field

[0001] The present application relates to the field of information retrieval technology, and in particular to a data analysis method, device, electronic device and computer program product. Background Art

[0002] When searching for information in a professional field, the generated answer data is often highly professional, which requires users to have high professional literacy to be able to interpret the generated answer data, resulting in poor readability of the answer data. Therefore, how to improve the readability of the answer data has become a technical problem that needs to be solved urgently by those skilled in the art. Summary of the invention

[0003] In view of this, the present application proposes a data analysis method, device, electronic device and computer program product, which can improve the readability of answer data.

[0004] The technical solutions proposed in this application are as follows:

[0005] In a first aspect, an embodiment of the present application provides a data analysis method, comprising:

[0006] Interpreting answer data determined according to the question query data through a pre-trained data interpretation model, and outputting interpretation information of the answer data;

[0007] The data interpretation model is obtained by inputting the knowledge of a specific industry into the industry vertical big model of the specific industry so that the industry vertical big model learns the knowledge of the specific industry; the knowledge of the specific industry includes knowledge fields, and also includes at least one of field meta-knowledge, field default knowledge and field sequence knowledge, the field meta-knowledge includes the knowledge interpretation of the knowledge field in the specific industry, the field default knowledge includes the common name of the knowledge field in the specific industry, and the field sequence knowledge includes the semantic order of the knowledge field; the specific industry includes the industry to which the question query data belongs.

[0008] Furthermore, the above method further includes:

[0009] Parsing the question query data, and determining a target knowledge element matching the question query data from a plurality of knowledge elements; the knowledge element is determined based on historical question query records, and the knowledge element is used to indicate a query target of the question query data;

[0010] A query statement is generated according to the question query data and the target knowledge element, and a query is performed in the data corresponding to the target knowledge element in the business database table according to the query statement to obtain answer data.

[0011] Furthermore, in the above method, the parsing of the question query data and determining a target knowledge element matching the question query data from a plurality of knowledge elements includes:

[0012] The question query data is input into a pre-trained data interpretation model so that the data interpretation model analyzes the question query data and determines a target knowledge element that matches the question query data from a plurality of knowledge elements.

[0013] Furthermore, the above method further includes:

[0014] Obtaining user feedback information for the target knowledge element;

[0015] The user feedback information is used to perform intensive training on the data interpretation model.

[0016] Furthermore, in the above method, generating a query statement according to the question query data and the target knowledge element includes:

[0017] The problem query data and the target knowledge element are converted into query codes through the data interpretation model, and the query codes are determined as the query statements.

[0018] Further, in the above method, the answer data includes a chart;

[0019] The step of querying the data corresponding to the target knowledge element in the business database table according to the query statement to obtain answer data includes:

[0020] According to the query statement, query the corresponding data content in the data corresponding to the target knowledge element in the business database table;

[0021] Generate a chart of a specified type according to the data content, and / or generate a general chart corresponding to the question query data.

[0022] Further, in the above method, searching for corresponding data content in the data corresponding to the target knowledge element in the business database table according to the query statement, generating a chart of a specified type according to the data content, and / or generating a general chart corresponding to the problem query data, includes:

[0023] Through the data interpretation model, according to the query statement, the corresponding data content is queried in the data corresponding to the target knowledge element in the business database table, a specified type of chart is generated according to the data content, and / or a general chart corresponding to the problem query data is generated.

[0024] In a second aspect, an embodiment of the present application provides a data analysis device, comprising:

[0025] An interpretation module, used to interpret answer data determined according to the question query data through a pre-trained data interpretation model, and output interpretation information of the answer data;

[0026] The data interpretation model is obtained by inputting the knowledge of a specific industry into the industry vertical big model of the specific industry so that the industry vertical big model learns the knowledge of the specific industry; the knowledge of the specific industry includes knowledge fields, and also includes at least one of field meta-knowledge, field default knowledge and field sequence knowledge, the field meta-knowledge includes the knowledge interpretation of the knowledge field in the specific industry, the field default knowledge includes the common name of the knowledge field in the specific industry, and the field sequence knowledge includes the semantic order of the knowledge field; the specific industry includes the industry to which the question query data belongs.

[0027] In a third aspect, an embodiment of the present application provides an electronic device, including:

[0028] A memory and a processor; wherein the memory is used to store programs; and the processor is used to implement any of the methods described above by running the programs in the memory.

[0029] In a fourth aspect, an embodiment of the present application provides a computer program product, the computer program product comprising a computer program, and when the computer program is executed by a processor, the computer program implements any one of the above methods. Optionally, the computer program can be stored in a readable storage medium of a computer device or in the cloud; the processor of the computer device reads the computer program from the readable storage medium or the cloud.

[0030] The data analysis method proposed in this application interprets the answer data determined according to the question query data through a pre-trained data interpretation model, and outputs the interpretation information of the answer data. The data interpretation model is obtained by inputting the knowledge of a specific industry into the industry vertical big model of the specific industry so that the industry vertical big model can learn the knowledge of the specific industry; the knowledge of the specific industry includes knowledge fields, and also includes at least one of field meta-knowledge, field default knowledge and field order knowledge, field meta-knowledge includes the knowledge interpretation of the knowledge field in the specific industry, field default knowledge includes the common name of the knowledge field in the specific industry, and field order knowledge includes the semantic order of the knowledge field; the specific industry includes the industry to which the question query data belongs. In this way, the data interpretation model obtained by knowledge learning based on the industry vertical big model can fully understand the data structure of the industry where the question query data is located, and then correctly and comprehensively interpret the answer data, thereby improving the readability of the answer data. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0032] Figure 1 It is a schematic diagram of a feasible application scenario of a data analysis method provided in an embodiment of the present application.

[0033] Figure 2 It is a flow chart of a data analysis method provided in an embodiment of the present application.

[0034] Figure 3 It is a schematic diagram of the process of retrieving answer data provided in an embodiment of the present application.

[0035] Figure 4 This is a schematic diagram of a page for users to select knowledge elements provided in an embodiment of the present application.

[0036] Figure 5 It is a flowchart of another data analysis method provided in an embodiment of the present application.

[0037] Figure 6 It is a schematic diagram of the output page of the data table provided in the embodiment of the present application.

[0038] Figure 7 This is a schematic diagram of the output page of the chart provided in the embodiment of the present application.

[0039] Figure 8 It is a schematic diagram of an output page for interpreting information provided in an embodiment of the present application.

[0040] Fig. 9 It is a structural schematic diagram of a data analysis device provided in an embodiment of the present application.

[0041] Fig.10 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions 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 creative work are within the scope of protection of this application.

[0043] When searching for information in a professional field, the generated answer data is usually rich in highly specialized content, which often includes the theoretical framework, industry terms, and complex analysis methods of the professional field. This requires users to have a high level of professional literacy, that is, if users want to accurately understand and effectively use this information, they must have the corresponding professional knowledge and background, which undoubtedly increases the difficulty for ordinary users to understand the information, resulting in poor readability of the answer data. Therefore, how to improve the readability of the answer data has become a technical problem that technicians in this field need to solve urgently.

[0044] Based on this, the present application proposes a data analysis method, device, electronic device and computer program product. The technical solution trains the data interpretation model to enable it to fully understand the data structure of the industry in which the question query data is located, and then correctly and comprehensively interpret the answer data, thereby achieving the effect of improving the readability of the answer data.

[0045] Figure 1 The feasible application scenarios of the data analysis method are shown, such as Figure 1 In the scenario shown, a client and a server are set up.

[0046] The above-mentioned vehicle refers to a motor vehicle, which may be a new energy vehicle or a conventional energy vehicle, and this embodiment does not limit this.

[0047] The client can be an electronic device with network access capability. Specifically, for example, the client can be a desktop computer, a tablet computer, a laptop computer, a smart phone, a digital assistant, a smart wearable device, a shopping guide terminal, a television, etc. Among them, the smart wearable device includes but is not limited to a smart bracelet, a smart watch, a smart glasses, a smart helmet, a smart necklace, etc. Alternatively, the client can also be software that can run in an electronic device.

[0048] The server can be an electronic device with certain computing and processing capabilities. It can have a network communication module, a processor, and a memory, etc. Of course, the server can also refer to software running in an electronic device. The server can also be a distributed server, which can be a system with multiple processors, memories, network communication modules, etc. operating in collaboration. Alternatively, the server can also be a server cluster formed by several servers. Alternatively, with the development of science and technology, the server can also be a new technical means that can realize the corresponding functions of the implementation method of the specification. For example, it can be a new form of "server" based on quantum computing.

[0049] The client and the server can communicate through the target network. The target network can be any type of network. For example, the target network can be a network that can be subdivided into multiple subnetworks. The target network or the multiple subnetworks contained in the target network can be at least one of a cellular mobile network (such as 2G, 3G, 4G or 5G), ZIGBEE, WIFI, Bluetooth, or any combination of at least one of these networks and other networks.

[0050] In the above feasible application scenario, the client is used to obtain question query data and send the question query data to the server. The server performs a search query based on the question query data to obtain the corresponding answer data, and then interprets the answer data determined according to the question query data through a pre-trained data interpretation model, and outputs interpretation information of the answer data. The server summarizes the answer data and the interpretation information of the answer data and sends them to the client, which displays them to the user.

[0051] It should be noted that the data interpretation model is obtained by inputting the knowledge of a specific industry into the industry vertical big model of the specific industry so that the industry vertical big model can learn the knowledge of the specific industry; the knowledge of a specific industry includes knowledge fields, and also includes at least one of field meta-knowledge, field default knowledge and field order knowledge, field meta-knowledge includes the knowledge interpretation of the knowledge field in the specific industry, field default knowledge includes the common name of the knowledge field in the specific industry, and field sequence knowledge includes the semantic order of the knowledge field; the specific industry includes the industry to which the question query data belongs.

[0052] With this setting, the data interpretation model obtained through knowledge learning based on the industry vertical big model can fully understand the data structure of the industry where the question query data is located, and then correctly and comprehensively interpret the answer data, thereby improving the readability of the answer data.

[0053] Furthermore, the present application embodiment proposes a data analysis method, which can be executed by an electronic device, which can be any device with data and instruction processing functions, for example, a notebook computer, a tablet computer, a desktop computer, a mobile device (for example, a mobile phone, a personal digital assistant, a dedicated messaging device) and other types of user terminals or a combination of any two or more of these electronic devices, or a server in the above embodiment. Figure 2 As shown, the method includes:

[0054] S101. Interpret answer data determined according to question query data through a pre-trained data interpretation model, and output interpretation information of the answer data.

[0055] The above-mentioned question query data refers to the question raised by the user, and the answer data refers to the data that answers the question. In some embodiments, only the question query data may be obtained, and a search may be performed based on the question query data to obtain the answer data that answers the question query data, so as to facilitate the interpretation of the answer data. In other embodiments, the answer data determined according to the question query data may be directly obtained, so as to facilitate the interpretation of the answer data.

[0056] In the embodiment of the present application, a data interpretation model is pre-trained, and the data interpretation model can interpret the answer data to obtain interpretation information of the answer data.

[0057] The data interpretation model is obtained after knowledge injection into the industry vertical big model. Among them, the industry vertical big model refers to a large language model (LLM, Large Language Model) specially customized for a specific industry or field. The industry vertical big model is trained through deep learning technology, and a large amount of specific industry text data is combined during the training process so that the industry vertical big model can better understand and generate natural language content related to the industry. Unlike general large-scale language models, industry vertical big models focus on a specific field, such as medical, legal, financial, education, etc., and they can provide more accurate and targeted services and support. Knowledge injection into the industry vertical big model can further enhance the professionalism of the industry vertical big model in the industry.

[0058] In the embodiment of the present application, according to the specific industry or field to which the data interpretation model is applied, the industry vertical big model of the specific industry is selected as the basic model, and then the knowledge of the specific industry is input into the basic model, so that the basic model learns the knowledge of the specific industry and obtains the data interpretation model, that is, the knowledge of the specific industry is used to inject knowledge into the industry vertical big model. It should be noted that the industry to which the question query data belongs should be the same industry as the specific industry to which the data interpretation model is applied.

[0059] For example, if the specific industry to which the data interpretation model is applied is the education industry, then the industry vertical model of the education industry is selected as the basic model, and then the knowledge of the education industry is input into the basic model so that the basic model can learn the knowledge of the education industry and obtain the data interpretation model. The data interpretation model can be used to interpret the answer data determined based on the query data related to the education industry, and output the interpretation information of the answer data.

[0060] Furthermore, the knowledge of a specific industry includes knowledge fields, and also includes at least one of field meta-knowledge, field default knowledge, and field sequence knowledge. The knowledge field refers to the specific knowledge content of a specific industry, while field meta-knowledge, field default knowledge, and field sequence knowledge are explanations of the knowledge field from different perspectives.

[0061] Specifically, field meta-knowledge includes the knowledge interpretation or meaning of the knowledge field in a specific industry. Inputting field meta-knowledge into the industry vertical big model for learning can enable the obtained data interpretation model to learn the meaning of the knowledge field of the specific industry in the specific industry. For example, the specific industry is the education industry, and the knowledge fields include "secondary vocational", "college" and "undergraduate". The field meta-knowledge of the knowledge field "secondary vocational" includes "minimum education level of primary school teachers"; the field meta-knowledge of the knowledge field "college" includes "minimum education level of junior high school teachers"; the field meta-knowledge of the knowledge field "undergraduate" includes "minimum education level of high school teachers". Based on this, the data interpretation model can further learn that "secondary vocational", "college" and "undergraduate" are all qualified education levels for primary school teachers, "college" and "undergraduate" are all qualified education levels for junior high school teachers, and "undergraduate" is a qualified education level for high school teachers.

[0062] Field default knowledge includes the common names of knowledge fields in specific industries. By learning the field default knowledge, when the question query data involves a common name, the data interpretation model can locate the knowledge field corresponding to the common name. For example, the specific industry is the education industry, and the knowledge field includes "education", and the field default knowledge of the knowledge field "education" includes "graduation information". When the question query data involves "graduation information", the data interpretation model can locate "education".

[0063] Field order knowledge includes the semantic order of knowledge fields. By learning field order knowledge, the data interpretation model can learn the implicit relationship between different knowledge fields. For example, if the specific industry is education, the knowledge fields include "doctoral students, master's students, undergraduate, junior college, general high school, secondary vocational", and the semantic order of knowledge fields includes "doctoral students > master's students > undergraduate > junior college > general high school > secondary vocational".

[0064] In a specific embodiment, the specific industry is the education industry, and the knowledge fields include {doctoral students, master's students, undergraduate students, junior college students, ordinary high schools, secondary vocational schools, and others}. The specific contents of the field meta-knowledge, field default knowledge, and field sequence knowledge are shown in Table 1.

[0065]

[0066] Table 1

[0067] In order to enable the data interpretation model to better perform the task of interpreting the answer data, training samples and training labels can also be constructed to perform supervised fine-tuning on the data interpretation model.

[0068] Specifically, a large amount of answer data can be obtained as training samples, and the interpretation information of the manually annotated answer data can be used as training labels. During fine-tuning, the training samples are input into the data interpretation model to obtain the inference results of the data interpretation model. By comparing the inference results of the data interpretation model with the training labels, the parameters of the data interpretation model are adjusted, and then the above fine-tuning process is repeated until the parameters of the data interpretation model meet the requirements.

[0069] In some embodiments, question query data may be obtained, and the data interpretation model may perform a search based on the question query data to obtain answer data that answers the question query data, interpret the answer data, and output the answer data and interpretation information of the answer data. In other embodiments, the answer data determined based on the question query data may be directly obtained, and the answer data may be interpreted using the data interpretation model to output interpretation information of the answer data.

[0070] In the above embodiment, the answer data determined according to the question query data is interpreted through a pre-trained data interpretation model, and the interpretation information of the answer data is output. The data interpretation model is obtained by inputting the knowledge of a specific industry into the industry vertical big model of the specific industry so that the industry vertical big model learns the knowledge of the specific industry; the knowledge of the specific industry includes knowledge fields, and also includes at least one of field meta-knowledge, field default knowledge and field sequence knowledge. In this way, the data interpretation model obtained by knowledge learning based on the industry vertical big model can fully understand the data structure of the industry where the question query data is located, and then correctly and comprehensively interpret the answer data, thereby improving the readability of the answer data.

[0071] As an optional implementation, Figure 3 As shown, in another embodiment of the present application, the method of the above embodiment may further include the following steps:

[0072] S201, parsing the question query data, and determining a target knowledge element matching the question query data from a plurality of knowledge elements.

[0073] The knowledge element is used to indicate the query target of the question query data, and one question query data can correspond to multiple knowledge elements. For example, if the question query data is "query teachers who graduated from school A", then the query target of the question query data can be determined to be "education experience: school name", and the knowledge element of the question query data can also be determined to be "education experience: school name".

[0074] The above-mentioned multiple knowledge elements are determined based on historical question query records. Historical question query records include the label configuration of query and filter components used by relevant users in historical time periods, etc. The user behavior points can be analyzed based on system data to extract the label configuration of query and filter components previously used by relevant users to determine the knowledge elements. In some other embodiments, the user behavior points can be analyzed based on system data to extract the label configuration of query and filter components previously used by relevant users to determine the query scenario and the knowledge elements in each scenario.

[0075] In a specific embodiment, the specific industry is the education industry. According to the system data analysis of user behavior points, the tag configuration of the query and screening components previously used by the teacher user is extracted to determine the teacher's basic information, education experience, work experience, personal awards, research topics and other scenarios and the knowledge elements in each scenario. Taking the teacher's education experience scenario as an example, some of the knowledge elements are shown in Table 2.

[0076] Scenario Knowledge Meta Teacher Education Experience Educational Background-School Name Teacher Education Experience Educational experience-school level Teacher Education Experience Educational Background-Major Teacher Education Experience Educational experience-education attributes Teacher Education Experience Educational Background-Graduation Date Teacher Education Experience Educational Background - Degree

[0077] Table 2

[0078] In an embodiment of the present application, the question query data is analyzed, and a target knowledge element corresponding to the question query data is selected from predetermined knowledge elements.

[0079] In some embodiments, the data interpretation model of the above embodiment is used to automatically match the target knowledge element corresponding to the question query data based on the question query data. Specifically, the question query data is input into the pre-trained data interpretation model so that the data interpretation model analyzes the question query data and determines the target knowledge element matching the question query data from multiple knowledge elements.

[0080] More specifically, after obtaining the question query data, the question query data is input into the data interpretation model, and the data interpretation model uses retrieval-augmented generation (RAG) technology to select target knowledge elements that match the question query data from multiple predetermined knowledge elements.

[0081] It should be noted that the number of target knowledge elements that match the question query data can be one or more, and the specific number can be set according to actual needs, and this embodiment does not limit it. The data interpretation model can select one or more knowledge elements with the highest similarity to the question query data from multiple predetermined knowledge elements as the target knowledge elements that match the question query data. For example, if the number of target knowledge elements is 10, the knowledge elements can be arranged from high to low according to the similarity with the question query data, and only the first 10 knowledge elements are retained as target knowledge elements, and the remaining knowledge elements are deleted.

[0082] If confusion occurs in the process of matching knowledge elements by the data interpretation model, such as the similarity of knowledge elements in scenarios with similar meanings such as work experience and part-time social jobs, it may cause confusion in knowledge element matching. In such a case, the user can manually specify the target knowledge that matches the question query data. For example, in the case of confusion in knowledge element matching, the user can choose to manually match knowledge elements. After obtaining the instruction to manually match knowledge elements, the knowledge elements corresponding to each scenario can be output for the user to choose. Figure 4 In the illustrated embodiment, the specific industry is the education industry, and after obtaining instructions for manually matching knowledge elements, the knowledge elements corresponding to the scenarios in each education industry are output.

[0083] When the data interpretation model is used to match knowledge elements, user feedback information for the target knowledge element can be obtained; the data interpretation model can be reinforced with the user feedback information. Specifically, the user feedback information for the target knowledge element includes positive information provided by the user when the target knowledge element is selected correctly, such as a thumbs-up, and negative information provided by the user when the target knowledge element is selected incorrectly, such as a thumbs-down. Reinforced training of the data interpretation model based on user feedback information can improve the accuracy of data analysis by the data interpretation model.

[0084] In addition, the user can also choose not to use the data interpretation model to match the knowledge element, but directly choose to manually match the knowledge element, which is not limited in this embodiment. Figure 4 Select knowledge element on the page shown.

[0085] Taking the teacher education experience scenario as an example, the knowledge elements corresponding to some question query data are shown in Table 3.

[0086] Scenario Question query data Knowledge Meta Teacher Education Experience Find teachers who graduated from University A Educational Background-School Name Teacher Education Experience Find teachers whose first degree is in basic mathematics Educational Background-Major Teacher Education Experience Screen out teachers who have obtained a master's degree on a part-time basis Educational experience-education attributes Teacher Education Experience Statistics of teachers graduating each year in the past five years Educational Background-Graduation Date Teacher Education Experience Search for teachers with a master's degree or doctorate Educational Background - Degree

[0087] Table 3

[0088] S202: Generate a query statement based on the question query data and the target knowledge element, and query the data corresponding to the target knowledge element in the business database table according to the query statement to obtain answer data.

[0089] The business database table stores relevant business data, including user information, project details, business information, etc. A search is performed in the business database table to determine the answer data of the question query data.

[0090] After obtaining the question query data and the target knowledge element, a query statement is generated according to the question query data and the target knowledge element, and a query code is generated according to the question query data and the target knowledge element as the query statement. In some embodiments, the question query data and the target knowledge element are converted into Python code as the query statement.

[0091] When storing business data, the business database table stores the business data according to different knowledge elements. When searching for question query data in the business database table, you only need to search the data under the target knowledge element corresponding to the question query data. For example, Table 4 is part of the content in the business database table under the education industry, which is stored according to knowledge elements such as age, gender, highest education, graduation school and degree. If the target knowledge element is the graduation school, you can search in the column where the graduation school knowledge element is located in Table 4. In this way, you can search for question query statements in a targeted manner without searching the entire business database table. The retrieval speed is fast and the reliability of the results is higher.

[0092] Name age gender Highest education Graduate School Bachelor of Science Little Red 25 female undergraduate School A Bachelor Xiao Ming 46 male undergraduate School B Bachelor

[0093] Table 4

[0094] Specifically, the user query data can be deconstructed in natural language based on the target knowledge element to obtain data analysis operations and data processing operations. Data analysis operations include operations that need to be performed to obtain answer data, including screening and counting operations performed on the data under the target knowledge element; data processing operations include processing operations on the data obtained by the data analysis operation, and the data obtained by the data analysis operation can be processed using the processing operation to obtain answer data. The data analysis operation and the data analysis operation results are encoded into query codes to obtain query statements.

[0095] In this way, by deconstructing the user query data based on the target knowledge element, a more specific query statement can be obtained, thus ensuring the accuracy of the result data.

[0096] For example, Table 5 shows the data analysis operations and data processing operations corresponding to different question query data under the teacher gender query knowledge element in the education industry.

[0097]

[0098] Table 5

[0099] In some embodiments, a code generation model can be trained, based on which user query data is deconstructed in natural language to obtain data analysis operations and data processing operations, and the data analysis operations and the results of the data analysis operations are encoded into query codes to obtain query statements. A large amount of user query data can be obtained as training samples, and the query codes corresponding to the training samples can be manually annotated as training labels, and the code generation model can be supervisedly trained using the training samples and training labels. The code generation model can use LLM as the basic model, which is not limited in this embodiment.

[0100] The data interpretation model of this embodiment not only uses the industry vertical model as the basic model, but also has learned the knowledge of a specific industry, so the data interpretation model can fully understand the data structure of the industry. If the data interpretation model is used to convert the query data and the target knowledge element into a query code, the reliability of the query code obtained will be higher.

[0101] In order to enable the data interpretation model to have the ability to convert question query data and target knowledge elements into query codes, training samples and labels can be constructed to fine-tune the data interpretation model in a supervised manner. For example, a large amount of user query data can be obtained as training samples, and the query codes corresponding to the training samples can be manually annotated as training labels. During fine-tuning, the training samples can be input into the data interpretation model to obtain the inference results of the data interpretation model. By comparing the inference results of the data interpretation model with the training labels, the loss of the data interpretation model can be determined. The parameters of the data interpretation model can be adjusted with the goal of reducing the loss of the data interpretation model, and then the above training process can be repeated until the parameters of the data interpretation model meet the requirements.

[0102] In this way, the problem query data and target knowledge elements can be converted into query codes through the data interpretation model, and the query codes can be determined as query statements.

[0103] Furthermore, the query code can be executed to query the business database table to obtain the answer data. It should be noted that different knowledge elements have their own corresponding scenarios, and the knowledge elements in the same scenario can be stored in the same library table, so that the business database table can contain library tables for multiple different scenarios. When querying in the business database table, you can first determine the library table where the target knowledge element is located, and then execute the query code for the library table after finding the library table to obtain the answer data. In this way, you do not need to query the entire business database table, and you can quickly locate the target knowledge element.

[0104] As an optional implementation, another embodiment of the present application discloses that the answer data in the above embodiment includes a chart, and the steps of the above embodiment perform a query in the data corresponding to the target knowledge element in the business database table according to the query statement to obtain the answer data, which may specifically include the following steps:

[0105] According to the query statement, the corresponding data content is queried in the data corresponding to the target knowledge element in the business database table; a specified type of chart is generated according to the data content, and / or a general chart corresponding to the problem query data is generated.

[0106] Charts can visualize data in a graphical manner, including bar charts, line charts, pie charts, etc. Charts can intuitively display trends, patterns, and relationships in data, making complex data easier to understand. In the embodiments of the present application, corresponding data is first retrieved from the data under the target knowledge element, and a chart can be generated using the retrieved data.

[0107] In some embodiments, if the user's question query statement specifies a chart type, such as a line chart or a bar chart, a chart of the specified type can be generated according to the data content. In other embodiments, if the user's question query statement does not specify a chart type, a suitable chart type and chart parameters can be determined according to the question query data, and then a general chart corresponding to the question query data can be generated.

[0108] In this way, a chart can be generated as the answer data output, so that the answer data can be displayed more clearly and intuitively.

[0109] It should be noted that the data content of the above embodiment includes a data table, which presents data in the form of rows and columns, and each cell of the data table contains specific numerical values ​​or text information, which is convenient for users to view the original data, compare and analyze. According to the user's personalized question query data, this embodiment can return the user's desired data table view for the user to view.

[0110] As an optional implementation, another embodiment of the present application discloses that the steps of the above embodiment include searching for corresponding data content in the data corresponding to the target knowledge element in the business database table according to the query statement, generating a chart of a specified type according to the data content, and / or generating a general chart corresponding to the problem query data, which may specifically include the following steps:

[0111] Through the data interpretation model, according to the query statement, the corresponding data content is queried in the data corresponding to the target knowledge element in the business database table, and a specified type of chart is generated according to the data content, and / or a general chart corresponding to the problem query data is generated.

[0112] The data interpretation model of this embodiment not only uses the industry vertical model as the basic model, but also has learned the knowledge of a specific industry, so the data interpretation model can fully understand the data structure of the industry. If the data interpretation model is used to generate charts, the reliability of the resulting charts will be higher.

[0113] In order to enable the data interpretation model to generate charts, it is necessary to fine-tune the data interpretation model. Specifically, a large amount of business data can be obtained as training labels, and then various charts of the business data can be manually drawn as training labels. In this way, the data interpretation model can be fine-tuned in a supervised manner using training samples and training labels, so that the data interpretation model has the ability to generate charts based on relevant data. At the same time, the general charts corresponding to different question query statements can be input into the data interpretation model as knowledge, so that the data interpretation model can automatically generate general charts based on the question query data when no chart type instructions are obtained.

[0114] The data interpretation model first executes the query statement, searches for the corresponding data content in the data corresponding to the target knowledge element in the business database table, and then generates a chart of a specified type based on the data content, and / or generates a general chart corresponding to the problem query data.

[0115] In this way, the data interpretation model can be used to quickly generate the required charts and graphs with higher reliability.

[0116] In a specific embodiment, the question query data may be input into a data interpretation model, and the data interpretation model may perform the following operations:

[0117] Parsing the question query data, and determining a target knowledge element matching the question query data from multiple knowledge elements;

[0118] Generate a query statement based on the question query data and the target knowledge element, and query the data corresponding to the target knowledge element in the business database table according to the query statement to obtain the answer data;

[0119] The answer data determined according to the question query data are interpreted, and interpretation information of the answer data is output; the answer data and the interpretation information of the answer data are output for the user to review.

[0120] The specific process of the data interpretation model performing the above operations can be referred to the records of the above embodiments and will not be elaborated here.

[0121] like Figure 5 As shown below, taking the education industry as an example, the data analysis method is explained in detail.

[0122] In advance, it is necessary to analyze the teacher user behavior points based on system data, extract the label configuration of the query and screening components previously used by the teacher user, determine the teacher's basic information, educational experience, work experience, personal awards, research topics and other scenarios and the knowledge elements in each scenario.

[0123] like Figure 5 As shown, the specific steps are as follows:

[0124] Obtain teacher data analysis instructions as query statements for educational industry issues;

[0125] Use the data interpretation model to match the knowledge element corresponding to the question query statement;

[0126] Determine whether the knowledge element matched by the data interpretation model is correct; if the knowledge element matched by the data interpretation model is incorrect, the user is required to manually input the knowledge element;

[0127] After obtaining the correct knowledge element, a data analysis code is generated as a query statement based on the knowledge element and the question query statement;

[0128] By executing the query statement, a data retrieval view is generated as a data table; for example, Figure 6 The data table shown is the corresponding data table generated when the question query data is "Query teachers who graduated from University A".

[0129] Generate a visual chart; for example, Figure 7 The following is the chart generated when the query data is "Count the number of awards won by teachers each year from 2021 to 2023, and generate a line chart".

[0130] Interpret data tables and charts to obtain interpretation information; for example, Figure 8 Shown is the interpretation information provided by an embodiment.

[0131] Corresponding to the above data analysis method, the present application embodiment also discloses a data analysis device, see Fig. 9 As shown, the device comprises:

[0132] The interpretation module 100 is used to interpret the answer data determined according to the question query data through a pre-trained data interpretation model, and output interpretation information of the answer data;

[0133] The data interpretation model is obtained by inputting the knowledge of a specific industry into the industry vertical big model of the specific industry so that the industry vertical big model can learn the knowledge of the specific industry; the knowledge of a specific industry includes knowledge fields, and also includes at least one of field meta-knowledge, field default knowledge and field sequence knowledge, field meta-knowledge includes the knowledge interpretation of the knowledge field in the specific industry, field default knowledge includes the common name of the knowledge field in the specific industry, and field sequence knowledge includes the semantic order of the knowledge field; the specific industry includes the industry to which the problem query data belongs.

[0134] As an optional implementation, another embodiment of the present application discloses that the device in the above embodiment may further include:

[0135] A parsing module is used to parse the question query data and determine a target knowledge element that matches the question query data from multiple knowledge elements; the knowledge element is determined based on historical question query records, and the knowledge element is used to indicate a query target of the question query data;

[0136] The query module is used to generate a query statement based on the question query data and the target knowledge element, and query the data corresponding to the target knowledge element in the business database table according to the query statement to obtain the answer data.

[0137] As an optional implementation method, another embodiment of the present application discloses that the parsing module of the above embodiment is specifically used to input the question query data into a pre-trained data interpretation model so that the data interpretation model parses the question query data and determines the target knowledge element that matches the question query data from multiple knowledge elements.

[0138] As an optional implementation, another embodiment of the present application discloses that the device in the above embodiment may further include:

[0139] The reinforcement module is used to obtain user feedback information on the target knowledge element; the user feedback information is used to perform reinforcement training on the data interpretation model.

[0140] As an optional implementation method, another embodiment of the present application discloses that the query module of the above embodiment is specifically used to convert the question query data and the target knowledge element into a query code through a data interpretation model, and determine the query code as a query statement.

[0141] As an optional implementation method, another embodiment of the present application discloses that the query module of the above embodiment is specifically used to query the corresponding data content in the data corresponding to the target knowledge element in the business database table according to the query statement; generate a specified type of chart according to the data content, and / or generate a general chart corresponding to the problem query data.

[0142] As an optional implementation method, another embodiment of the present application discloses that the query module of the above embodiment is specifically used to query the corresponding data content in the data corresponding to the target knowledge element in the business database table according to the query statement through the data interpretation model, generate a specified type of chart according to the data content, and / or generate a general chart corresponding to the problem query data.

[0143] Specifically, the device provided in this embodiment belongs to the same application concept as the method provided in the above embodiment of this application, can execute the method provided in any of the above embodiments of this application, and has the corresponding functional modules and beneficial effects of the execution method. For the technical details not fully described in this embodiment, please refer to the specific processing content of the method provided in the above embodiment of this application, which will not be repeated here.

[0144] The functions implemented by the above modules can be implemented by the same or different processors respectively, and the embodiments of the present application are not limited thereto.

[0145] It should be understood that the modules in the above devices can be implemented in the form of a processor calling software. For example, the device includes a processor, the processor is connected to a memory, and instructions are stored in the memory. The processor calls the instructions stored in the memory to implement any of the above methods or realize the functions of each module of the device, wherein the processor can be a general-purpose processor, such as a CPU or a microprocessor, etc., and the memory can be a memory in the device or a memory outside the device. Alternatively, the modules in the device can be implemented in the form of a hardware circuit, and the functions of some or all modules can be realized by designing the hardware circuit. The hardware circuit can be understood as one or more processors; for example, in one implementation, the hardware circuit is an ASIC, and the functions of some or all modules above are realized by designing the logical relationship of the components in the circuit; for another example, in another implementation, the hardware circuit can be implemented by PLD, taking FPGA as an example, which can include a large number of logic gate circuits, and the connection relationship between the logic gate circuits is configured by a configuration file, so as to realize the functions of some or all modules above. All modules of the above devices can be implemented in the form of a processor calling software, or in the form of a hardware circuit, or in the form of a processor calling software, and the remaining part can be implemented in the form of a hardware circuit.

[0146] In an embodiment of the present application, a processor is a circuit with the ability to process signals. In one implementation, the processor may be a circuit with the ability to read and run instructions, such as a CPU, a microprocessor, a GPU, or a DSP; in another implementation, the processor may implement certain functions through the logical relationship of a hardware circuit, and the logical relationship of the hardware circuit is fixed or reconfigurable, such as a hardware circuit implemented by an ASIC or PLD, such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration document to implement the hardware circuit configuration can be understood as the process of the processor loading instructions to implement the functions of some or all of the above modules. In addition, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as an NPU, TPU, DPU, etc.

[0147] It can be seen that each module in the above device can be one or more processors (or processing circuits) configured to implement the above method, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.

[0148] In addition, the modules in the above device can be fully or partially integrated together, or can be implemented independently. In one implementation, these modules are integrated together and implemented in the form of a SOC. The SOC may include at least one processor for implementing any of the above methods or implementing the functions of the modules of the device. The type of the at least one processor may be different, for example, including a CPU and an FPGA, a CPU and an artificial intelligence processor, a CPU and a GPU, etc.

[0149] An embodiment of the present application further provides a control device, which includes a processor and an interface circuit. The processor in the control device is connected to an input-output component through the interface circuit of the control device.

[0150] The input-output component specifically refers to a hardware component that enables a user to input information and output information to the user, such as a microphone, keyboard, handwriting tablet, touch screen, display, speaker, printer, etc.

[0151] The above-mentioned interface circuit can be any interface circuit that can realize the data communication function, for example, it can be a USB interface circuit, a Type-C interface circuit, a serial port circuit, a PCIE circuit, etc.

[0152] The processor in the control device is a circuit with signal processing capability, which can improve the readability of the answer data by executing any of the data analysis methods introduced in the above embodiments. The specific implementation of the processor can refer to the above processor implementation, and the embodiments of this application are not strictly limited.

[0153] When the control device is applied to a device with a human-computer interaction function, the input and output components of the control device may be input components and output components on the device, such as a microphone, a keyboard, a handwriting tablet, a touch screen, a display, an audio player, etc. At the same time, the processor of the control device may be a CPU or GPU, etc. provided by the device, and the interface circuit of the control device may be an interface circuit between the information input component of the device and a processor such as a CPU or GPU.

[0154] Corresponding to the above data analysis method, the present application embodiment also discloses an electronic device, see Fig.10 As shown, the electronic device includes:

[0155] Memory 200 and processor 210;

[0156] The memory 200 is connected to the processor 210 and is used to store programs;

[0157] The processor 210 is used to implement the data analysis method disclosed in any of the above embodiments by running the program stored in the memory 200.

[0158] Specifically, the electronic device may further include: a bus, a communication interface 220 , an input device 230 and an output device 240 .

[0159] The processor 210, the memory 200, the communication interface 220, the input device 230 and the output device 240 are connected to each other via a bus.

[0160] A bus may include a pathway that transfers information between components of a computer system.

[0161] The processor 210 may be a general-purpose processor, such as a general-purpose central processing unit (CPU), a microprocessor, etc., or an application-specific integrated circuit (ASIC), or one or more integrated circuits for controlling the execution of the program of the present application. It may also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0162] The processor 210 may include a main processor, and may also include a baseband chip, a modem, and the like.

[0163] The memory 200 stores a program for executing the technical solution of the present application, and may also store an operating system and other key services. Specifically, the program may include a program code, and the program code includes computer operation instructions. More specifically, the memory 200 may include a read-only memory (ROM), other types of static storage devices that can store static information and instructions, a random access memory (RAM), other types of dynamic storage devices that can store information and instructions, a disk storage, a flash, and the like.

[0164] The input device 230 may include a device for receiving data and information input by a user, such as a keyboard, a mouse, a camera, a scanner, a light pen, a voice input device, a touch screen, a pedometer, or a gravity sensor.

[0165] Output device 240 may include devices that allow information to be output to a user, such as a display screen, printer, speaker, etc.

[0166] The communication interface 220 may include any transceiver or the like to communicate with other devices or communication networks, such as Ethernet, Radio Access Network (RAN), Wireless Local Area Network (WLAN), etc.

[0167] The processor 210 executes the program stored in the memory 200 and calls other devices, which can be used to implement the various steps of the data analysis method provided in the above embodiments of the present application.

[0168] In addition to the above methods and devices, the embodiments of the present application may also be computer program products, which include computer programs. When the computer programs are executed by a processor, the data analysis methods provided by any of the above embodiments of the present application may be executed. Optionally, the computer program may be stored in a readable storage medium or in the cloud of a computer device; the processor of the computer device reads the computer program from the readable storage medium or the cloud.

[0169] The computer program product may be written in any combination of one or more programming languages ​​to write program codes for performing the operations of the embodiments of the present application, including object-oriented programming languages ​​such as Java, C++, etc., and also conventional procedural programming languages ​​such as "C" language or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0170] The computer program product may be implemented in hardware, software or a combination thereof. In one optional embodiment, the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (SDK).

[0171] In addition, an embodiment of the present application may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes each step of the data analysis method provided in the above embodiment.

[0172] The computer readable storage medium may adopt any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may include, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an 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 above.

[0173] For the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the order of the actions described, because according to the present application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for the present application.

[0174] It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0175] The steps in the methods of each embodiment of the present application can be adjusted in order, combined and deleted according to actual needs, and the technical features recorded in each embodiment can be replaced or combined.

[0176] The modules and sub-modules in the devices and terminals in the various embodiments of the present application can be combined, divided and deleted according to actual needs.

[0177] In the several embodiments provided in the present application, it should be understood that the disclosed terminals, devices and methods can be implemented in other ways. For example, the terminal embodiments described above are only schematic, for example, the division of modules or submodules is only a logical function division, and there may be other division methods in actual implementation, such as multiple submodules or modules can be combined or integrated into another module, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0178] The modules or submodules described as separate components may or may not be physically separated, and the components of the modules or submodules may or may not be physical modules or submodules, that is, they may be located in one place, or they may be distributed on multiple network modules or submodules. Some or all of the modules or submodules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0179] In addition, each functional module or submodule in each embodiment of the present application may be integrated into one processing module, or each module or submodule may exist physically separately, or two or more modules or submodules may be integrated into one module. The above-mentioned integrated modules or submodules may be implemented in the form of hardware or in the form of software functional modules or submodules.

[0180] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0181] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly by hardware, software units executed by a processor, or a combination of the two. The software units may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0182] Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "including a..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0183] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A data analysis method, characterized in that: include: Interpreting answer data determined according to the question query data through a pre-trained data interpretation model, and outputting interpretation information of the answer data; The data interpretation model is obtained by inputting the knowledge of a specific industry into the industry vertical big model of the specific industry so that the industry vertical big model learns the knowledge of the specific industry; the knowledge of the specific industry includes knowledge fields, and also includes at least one of field meta-knowledge, field default knowledge and field sequence knowledge, the field meta-knowledge includes the knowledge interpretation of the knowledge field in the specific industry, the field default knowledge includes the common name of the knowledge field in the specific industry, and the field sequence knowledge includes the semantic order of the knowledge field; the specific industry includes the industry to which the question query data belongs.

2. The method according to claim 1, characterized in that Also includes: Parsing the question query data, and determining a target knowledge element matching the question query data from a plurality of knowledge elements; The knowledge element is determined based on historical question query records, and the knowledge element is used to indicate a query target of the question query data; A query statement is generated according to the question query data and the target knowledge element, and a query is performed in the data corresponding to the target knowledge element in the business database table according to the query statement to obtain answer data.

3. The method according to claim 2, characterized in that The step of parsing the question query data and determining a target knowledge element matching the question query data from a plurality of knowledge elements includes: The question query data is input into a pre-trained data interpretation model so that the data interpretation model analyzes the question query data and determines a target knowledge element that matches the question query data from a plurality of knowledge elements.

4. The method according to claim 3, characterized in that Also includes: Obtaining user feedback information for the target knowledge element; The user feedback information is used to perform intensive training on the data interpretation model.

5. The method according to claim 2, characterized in that: The step of generating a query statement according to the question query data and the target knowledge element includes: The problem query data and the target knowledge element are converted into query codes through the data interpretation model, and the query codes are determined as the query statements.

6. The method according to claim 2, characterized in that The answer data includes a chart; The step of querying the data corresponding to the target knowledge element in the business database table according to the query statement to obtain answer data includes: According to the query statement, query the corresponding data content in the data corresponding to the target knowledge element in the business database table; Generate a chart of a specified type according to the data content, and / or generate a general chart corresponding to the question query data.

7. The method according to claim 6, characterized in that The step of searching for corresponding data content in the data corresponding to the target knowledge element in the business database table according to the query statement, generating a chart of a specified type according to the data content, and / or generating a general chart corresponding to the problem query data includes: Through the data interpretation model, according to the query statement, the corresponding data content is queried in the data corresponding to the target knowledge element in the business database table, a specified type of chart is generated according to the data content, and / or a general chart corresponding to the problem query data is generated.

8. A data analysis device, characterized in that: include: An interpretation module, used to interpret answer data determined according to the question query data through a pre-trained data interpretation model, and output interpretation information of the answer data; The data interpretation model is obtained by inputting the knowledge of a specific industry into the industry vertical big model of the specific industry so that the industry vertical big model learns the knowledge of the specific industry; the knowledge of the specific industry includes knowledge fields, and also includes at least one of field meta-knowledge, field default knowledge and field sequence knowledge, the field meta-knowledge includes the knowledge interpretation of the knowledge field in the specific industry, the field default knowledge includes the common name of the knowledge field in the specific industry, and the field sequence knowledge includes the semantic order of the knowledge field; the specific industry includes the industry to which the question query data belongs.

9. An electronic device, characterized in that: include: Memory and processor; Wherein, the memory is used to store programs; The processor is used to implement the method according to any one of claims 1 to 7 by running the program in the memory.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images