Data analysis method and device, electronic equipment and computer readable medium

Abstract

The invention discloses a data analysis method and device, electronic equipment and a computer readable medium, and relates to the technical field of computers.The specific implementation mode comprises the steps that in response to a data analysis request, a corresponding post role identifier and demand text data are obtained; segmenting the required text data based on paragraph boundaries to obtain corresponding text blocks, inputting the text blocks into the embedded model to obtain corresponding text vectors, performing meta-information extraction on the text blocks to obtain corresponding meta-information, and inputting the text blocks into the large model to obtain corresponding keywords; calling a domain knowledge base corresponding to the post role identifier, so as to obtain a corresponding cue word based on the text vector, the meta-information and the keyword; and generating a data analysis result based on the cue word. Therefore, data sharing collaboration is achieved, the service process is effectively optimized, the data analysis efficiency and success rate are improved, and the basic level management work efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and particularly to a data analysis method, apparatus, electronic device, and computer-readable medium. Background Art

[0002] Currently, the applications of large models in vertical fields are mainly divided into two categories: domain knowledge base question answering and data analysis. The technical path of domain knowledge base question answering is to first construct a knowledge base using domain knowledge materials. When a user asks a question, relevant information fragments are recalled from the knowledge base through semantic retrieval, and then these information fragments are combined with the user's question to form a direct input prompt for the large model and passed to the large model to obtain an answer. The technical path of constructing a domain knowledge base includes parsing documents, segmenting the documents according to rules, and converting the segmented document fragments into vectors and storing them in a vector library. The technical path of data analysis is to first generate an SQL query statement based on the user's input and table schema information, and then execute the SQL statement to obtain the corresponding result. At present, assistant tools such as domain knowledge base question answering and data analysis are scattered, and the process links are fragmented, and knowledge sharing and collaboration are impossible. The current isolated assistant tools cannot effectively optimize the business process, resulting in low work efficiency of grass-roots managers. Summary of the Invention

[0003] In view of this, embodiments of this application provide a data analysis method, apparatus, electronic device, and computer-readable medium, which can solve the problems that assistant tools such as domain knowledge base question answering and data analysis are scattered at the current stage and cannot effectively optimize the business process, resulting in low work efficiency of grass-roots managers.

[0004] To achieve the above object, according to one aspect of the embodiments of this application, a data analysis method is provided, including:

[0005] Responding to a data analysis request, obtaining the corresponding job role identifier and requirement text data;

[0006] Segmenting the requirement text data based on paragraph boundaries to obtain corresponding text blocks, inputting the text blocks into an embedding model to obtain corresponding text vectors, extracting meta-information from the text blocks to obtain corresponding meta-information, and inputting the text blocks into a large model to obtain corresponding keywords;

[0007] Invoking the domain knowledge base corresponding to the job role identifier to query and obtain corresponding prompt words based on the text vectors, meta-information, and keywords;

[0008] Generating a data analysis result based on the prompt words.

[0009] Optionally, before invoking the domain knowledge base corresponding to the job role identifier, the method further includes:

[0010] Obtain the list of position identifiers corresponding to the preset business, and obtain the corresponding position knowledge data based on the list of position identifiers;

[0011] Convert the position knowledge data into text data, and then call the semantic segmentation model to perform text segmentation on the text data based on paragraph boundaries to obtain each text block;

[0012] Input each text block into the embedding model to obtain each text vector, extract meta-information for each text block to obtain each meta-information, and input each text block into the large model to obtain the corresponding keywords;

[0013] Obtain the corresponding risk response measures and corresponding prompt words for each position identifier in the list of position identifiers;

[0014] Build a domain knowledge base based on each text vector, each meta-information, each keyword, each prompt word, and each risk response strategy.

[0015] Optionally, before obtaining the corresponding position role identifier and demand text data, the method further includes:

[0016] Obtain business operation data in real time, call the anomaly monitoring program to monitor in real time whether the business operation data is abnormal, and if so, trigger a data analysis request.

[0017] Optionally, generate a data analysis result based on the prompt word, including:

[0018] Obtain the risk response strategy corresponding to the prompt word from the domain knowledge base as the data analysis result.

[0019] Optionally, generate a data analysis result according to the prompt word, including:

[0020] Determine the corresponding process link according to the prompt word;

[0021] Generate the corresponding data analysis result based on the process link.

[0022] Optionally, generate a data analysis result according to the prompt word, including:

[0023] In response to the process link including model prediction, call the optimal hyperparameter combination model to perform data prediction based on the text vector, meta-information, and keyword to obtain the predicted data analysis result.

[0024] In addition, the present application also provides a data analysis device, including:

[0025] An acquisition unit configured to obtain the corresponding position role identifier and demand text data in response to a data analysis request;

[0026] A keyword determination unit, configured to segment the requirement text data based on paragraph boundaries to obtain corresponding text blocks, input the text blocks into an embedding model to obtain corresponding text vectors, extract meta-information from the text blocks to obtain corresponding meta-information, and input the text blocks into a large model to obtain corresponding keywords;

[0027] A query unit, configured to call the domain knowledge base corresponding to the job role identifier to query and obtain corresponding prompt words based on the text vector, meta-information, and keywords;

[0028] A data analysis unit, configured to generate a data analysis result based on the prompt words.

[0029] Optionally, the device further includes a domain knowledge base construction unit, configured to:

[0030] Obtain a list of job identifiers corresponding to a preset business, and obtain corresponding job knowledge data based on the list of job identifiers;

[0031] Convert the job knowledge data into text data, and then call a semantic segmentation model to perform text segmentation on the text data based on paragraph boundaries to obtain each text block;

[0032] Input each text block into the embedding model to obtain each text vector, extract meta-information from each text block to obtain each meta-information, and input each text block into the large model to obtain corresponding keywords;

[0033] Obtain various risk response measures and corresponding prompt words corresponding to each job identifier in the list of job identifiers;

[0034] Construct a domain knowledge base based on each text vector, each meta-information, each keyword, each prompt word, and each risk response strategy.

[0035] Optionally, the device further includes an anomaly monitoring unit, configured to:

[0036] Obtain business operation data in real time, and call an anomaly monitoring program to monitor in real time whether the business operation data is abnormal. If so, trigger a data analysis request.

[0037] Optionally, the data analysis unit is further configured to:

[0038] Obtain a risk response strategy corresponding to the prompt word from the domain knowledge base as the data analysis result.

[0039] Optionally, the data analysis unit is further configured to:

[0040] Determine a corresponding process link according to the prompt word;

[0041] Generate corresponding data analysis results based on process steps.

[0042] Optionally, the data analysis unit is further configured to:

[0043] In response to the process step being configured as model prediction, call the optimal hyperparameter combination model to perform data prediction based on text vectors, meta-information, and keywords to obtain predicted data analysis results.

[0044] In addition, the present application also provides a data analysis electronic device, including: one or more processors; a storage device for storing one or more programs, which when executed by the one or more processors, cause the one or more processors to implement the data analysis method as described above.

[0045] In addition, the present application also provides a computer-readable medium having a computer program stored thereon, and when the program is executed by a processor, the data analysis method as described above is implemented.

[0046] One embodiment of the above invention has the following advantages or beneficial effects: In response to a data analysis request, the present application obtains the corresponding job role identifier and requirement text data; splits the requirement text data based on paragraph boundaries to obtain corresponding text blocks, inputs the text blocks into an embedding model to obtain corresponding text vectors, extracts meta-information from the text blocks to obtain corresponding meta-information, and inputs the text blocks into a large model to obtain corresponding keywords; calls the domain knowledge base corresponding to the job role identifier to query for corresponding prompt words based on the text vectors, meta-information, and keywords; and generates data analysis results based on the prompt words. Thereby, data sharing and collaboration are realized, the business process is effectively optimized, the data analysis efficiency and success rate are improved, and the grass-roots management work efficiency is also improved.

[0047] The further effects of the above non-conventional optional methods will be described in conjunction with specific embodiments below. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The drawings are used to better understand the present application and do not constitute an improper limitation to the present application. Among them:

[0049] Figure 1 is a schematic diagram of the main process of the data analysis method provided by an embodiment of the present application;

[0050] Figure 2 is a schematic diagram of the main process of the data analysis method provided by an embodiment of the present application;

[0051] Figure 3 is a schematic diagram of the principle of the data analysis method provided by an embodiment of the present application;

[0052] Figure 4 It is a schematic diagram of the process for constructing a domain knowledge base of the data analysis method provided by an embodiment of the present application;

[0053] Figure 5 It is a schematic diagram of the knowledge Q&A process of the data analysis method provided by an embodiment of the present application;

[0054] Figure 6 It is a schematic diagram of the main process of the data analysis method provided by an embodiment of the present application;

[0055] Figure 7 It is a schematic diagram of the main process of the data analysis method provided by an embodiment of the present application;

[0056] Figure 8 It is a schematic diagram of the main process of the data analysis method provided by an embodiment of the present application;

[0057] Figure 9 It is a schematic diagram of the main units of the data analysis device according to an embodiment of the present application;

[0058] Figure 10 It is an exemplary system architecture diagram to which the embodiments of the present application can be applied;

[0059] Figure 11 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing the embodiments of the present application. Detailed implementation manners

[0060] The following describes exemplary embodiments of the present application with reference to the accompanying drawings. Various details of the embodiments of the present application are included to assist understanding, and they should be considered merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted below. It should be noted that in the technical solutions of the present disclosure, in terms of the collection, gathering, updating, analysis, processing, use, transmission, storage, etc. of user personal information, they all comply with the provisions of relevant laws and regulations, are used for legal purposes, and do not violate public order and good customs. Necessary measures are taken for user personal information to prevent illegal access to user personal information data and to safeguard user personal information security, network security, and national security.

[0061] Figure 1 It is a schematic diagram of the main process of the data analysis method provided by an embodiment of the present application. As Figure 1 shown, the data analysis method includes:

[0062] Step S101, in response to a data analysis request, obtain the corresponding job role identifier and requirement text data.

[0063] In this embodiment, the execution entity of the data analysis method (for example, it can be a server) can receive a data analysis request through a wired connection or a wireless connection. Specifically, it can be a request for analyzing a user's question and giving a solution (i.e., the data analysis result). After receiving the data analysis request, the execution entity can obtain the job role identifier and requirement text data carried in the request.

[0064] Exemplarily, the requirement text data can include "grassroots managers need to obtain current warehousing personnel information" or "grassroots managers need to know the outbound volume of a certain category in the next 10 days".

[0065] Specifically, before obtaining the corresponding job role identifier and requirement text data, the data analysis method further includes: obtaining business operation data in real time, calling an anomaly monitoring program to monitor in real time whether the business operation data has anomalies, and if so, triggering a data analysis request.

[0066] In the logistics supply chain warehousing scenario, grassroots managers conduct data analysis on business operation data every day, squeeze information in real time, and judge whether there is a timeout risk based on preset conditions. When there is a timeout risk, an anomaly occurs and data analysis is triggered. Then the execution entity can query the corresponding domain knowledge base to obtain risk response measures to resolve the risk in a timely manner and ensure the normal operation of the logistics supply chain.

[0067] Step S102, segment the requirement text data based on paragraph boundaries to obtain corresponding text blocks, input the text blocks into an embedding model to obtain corresponding text vectors, extract meta-information from the text blocks to obtain corresponding meta-information, and input the text blocks into a large model to obtain corresponding keywords.

[0068] Use a semantic segmentation model to segment the requirement text data based on paragraph boundaries to obtain corresponding text blocks, thereby ensuring the integrity of context information. By inputting the text blocks into an embedding model to obtain corresponding text vectors, subsequent data analysis can be achieved more quickly and accurately.

[0069] Meta-information is information about information, used to describe the structure, semantics, usage, and application of information, etc. By extracting the meta-information corresponding to the text blocks, a more detailed understanding of the text blocks can be achieved, facilitating accurate subsequent data analysis.

[0070] A large model, in the field of artificial intelligence, refers to a deep neural network with more than 1 billion parameters, capable of processing massive amounts of data and completing various complex tasks, such as natural language processing, computer vision, speech recognition, etc.

[0071] By inputting a text block into a large model, corresponding keywords can be output. Each decision tree in the large model can calculate the importance scores of each word in the text block, and obtain the keywords from the words according to the importance scores.

[0072] By vectorizing the text block, extracting the original information, and obtaining keywords to achieve hierarchical data analysis, the results of data analysis can be made more accurate.

[0073] Step S103, call the domain knowledge base corresponding to the job role identifier to query and obtain the corresponding prompt words based on the text vector, meta-information, and keywords.

[0074] Each job role identifier has corresponding job knowledge materials, corresponding business database permissions, corresponding work scenarios, and corresponding work content for the corresponding work scenarios. Thus, the domain knowledge base corresponding to each job role identifier can be obtained. The execution entity can call the domain knowledge base corresponding to the job role identifier to find the prompt words corresponding to the text vector, meta-information, and keywords in the called domain knowledge base. The prompt words can be used to represent the pointer of the strategy for processing the questions raised by the user or the summary of the processing solution. The embodiments of the present application do not make specific limitations on the prompt words.

[0075] Step S104, generate the data analysis result based on the prompt words.

[0076] Specifically, generating the data analysis result based on the prompt words includes: obtaining the risk response strategy corresponding to the prompt words from the domain knowledge base as the data analysis result.

[0077] Taking the prompt words as pointers, the execution entity can index to the corresponding risk response strategy according to the prompt words, and use the obtained solution as the data analysis result.

[0078] In this embodiment, in response to a data analysis request, the corresponding job role identifier and demand text data are obtained; the demand text data is segmented based on the paragraph boundary to obtain the corresponding text blocks, the text blocks are input into an embedding model to obtain the corresponding text vectors, the meta-information of the text blocks is extracted to obtain the corresponding meta-information, the text blocks are input into a large model to obtain the corresponding keywords; the domain knowledge base corresponding to the job role identifier is called to query and obtain the corresponding prompt words based on the text vector, meta-information, and keywords; and the data analysis result is generated based on the prompt words. Thereby, data sharing and collaboration are realized, the business process is effectively optimized, the data analysis efficiency and success rate are improved, and the grass-roots management work efficiency is improved.

[0079] Figure 2 is the main process schematic diagram of the data analysis method provided by an embodiment of the present application, as Figure 2As shown in the figure, the data analysis method includes:

[0080] Step S201: Obtain the list of position identifiers corresponding to the preset business, and obtain the corresponding position knowledge data based on the list of position identifiers.

[0081] For each preset business, construct a corresponding domain knowledge base. Specifically, collect the position knowledge data corresponding to each position identifier in the list of position identifiers for each preset business. By way of example, in the process of obtaining the position knowledge data, the executing entity can abstract the position roles, summarize the position knowledge materials, the permissions of the business database, the corresponding work scenarios, and the work content corresponding to each work scenario, so as to summarize the corresponding position knowledge data.

[0082] Step S202: Convert the position knowledge data into text data, and then call the semantic segmentation model to perform text segmentation on the text data based on the paragraph boundaries to obtain each text block.

[0083] Step S203: Input each text block into the embedding model to obtain each text vector, extract meta-information from each text block to obtain each meta-information, and input each text block into the large model to obtain the corresponding keywords.

[0084] Step S204: Obtain the corresponding risk response measures and corresponding prompt words for each position identifier in the list of position identifiers.

[0085] Step S205: Construct a domain knowledge base based on each text vector, each meta-information, each keyword, each prompt word, and each risk response strategy.

[0086] Construct a domain knowledge base based on the position knowledge data. In order to improve the answering effect of the large model, in the stage of constructing the text knowledge base, the existing document segmentation method based on characters and word counts is improved, and the semantic segmentation model is used to segment the document according to the paragraph boundaries to ensure the integrity of the context information. Meta-information is information about information, which is used to describe the structure, semantics, uses, and usage methods of information.

[0087] By way of example, as Figure 4 shown, the position knowledge data in the local document includes unstructured data and structured data. Input the unstructured data and structured data in the local document into the loader to obtain the corresponding text, and then segment the text to obtain text blocks. Input each text block into the large language model (i.e., the LLM large model) to obtain keywords, input the text block into the embedding model to obtain the vector representation, extract the meta-information in the text block to obtain the corresponding meta-information, and then construct the vearch database, that is, the domain knowledge base, from the vector representation, meta-information, and keywords.

[0088] The constructed domain knowledge base can assist data analysis based on the association relationship between keywords, vector representations, and meta-information, and accurately and quickly obtain the result data required for data analysis.

[0089] Step S206: In response to a data analysis request, obtain the corresponding job role identifier and requirement text data.

[0090] Step S207: Segment the requirement text data based on paragraph boundaries to obtain corresponding text blocks. Input the text blocks into an embedding model to obtain corresponding text vectors, extract meta-information from the text blocks to obtain corresponding meta-information, and input the text blocks into a large model to obtain corresponding keywords.

[0091] Step S208: Invoke the domain knowledge base corresponding to the job role identifier to query for corresponding prompt words based on the text vectors, meta-information, and keywords.

[0092] Step S209: Generate a data analysis result based on the prompt words.

[0093] When receiving a data analysis request, as Figure 5 shown, the execution entity can input the requested requirement text data into an LLM model to obtain keywords, input the requested requirement text data into an embedding model to obtain corresponding vector representations, and extract the meta-information of the requirement text data. Then, based on the constructed domain knowledge base (Vearch database), perform meta-information query, keyword query, and vector query, and finally query for prompt words. Input the prompt words into the LLM model and the embedding model to obtain the final data analysis result data. In the knowledge retrieval stage, adopt a multi-channel recall method of meta-information + keyword + semantic vector to avoid the occurrence of domain misalignment.

[0094] Specifically, generating a data analysis result based on the prompt words includes: determining the corresponding process link according to the prompt words; generating the corresponding data analysis result based on the process link.

[0095] A complete data science process is as follows. Different work scenarios can correspond to some or all of the links. For example, in the supply chain warehousing scenario, if a grass-roots manager needs to obtain current warehousing personnel information, only the two links of data acquisition and result visualization in the process as Figure 6 shown are required. However, if a grass-roots manager needs to know the outbound volume of a certain category in the next 10 days, all the links in the process as Figure 7 shown are required. As Figure 7 shown, the data prediction process includes determining label fields, dataset splitting, model selection, hyperparameter automatic tuning, and model prediction.

[0096] Specifically, according to the prompt words, data analysis results are generated, including: in response to the process link including model prediction, the optimal hyperparameter combination model is called to perform data prediction based on text vectors, meta-information, and keywords to obtain the predicted data analysis results.

[0097] In the process as Figure 7 shown: The label determination link is to determine the feature information and label fields; the dataset splitting link is to determine the training set and test set to verify the model effect; the model selection stage is the lightgbm algorithm and the Prophet algorithm. Among them, the lightgbm algorithm can handle both regression problems and classification problems and has particularly good effects on tabular data. The Prophet algorithm is mainly used for sequence modeling, such as predicting the trend of a certain future time period based on historical data. Hyperparameter automatic tuning refers to using Bayesian thinking to determine the optimal hyperparameter combination in the hyperparameter search space. For regression, the MSE metric is used, and for classification tasks, the accuracy metric is used to evaluate the quality of the hyperparameter combination; the model prediction stage is to perform data prediction based on the optimal hyperparameter combination model.

[0098] Figure 8 is the main process schematic diagram of the data analysis method provided by an embodiment of the present application. As Figure 8 shown, for example, employees have accumulated learning in the knowledge base, database, and business system. The role-assisted copilot based on the LLM large model has undergone the same learning in the knowledge base, database, and business system. The real role interacts with the UI interface. The role-assisted copilot calls and interacts with the UI interface through the auxiliary interface and interacts with the server through the background call. The UI interface that interacts with the real role can also interact with the server. The role-assisted copilot can achieve knowledge sharing and collaboration with other real roles and other role-assisted copilots through interaction with the server, improving data analysis efficiency and accuracy. The role-assisted copilot can contain knowledge bases in various fields, facilitating and accurately querying strategies related to the prompt words, and this strategy can be the data analysis result.

[0099] Users can conveniently use natural language to simultaneously call the knowledge agent and the data role-assisted copilot, integrating them on the interface, which not only teaches knowledge, assists in simplifying offline processes, but also quickly obtains the results of business data fluctuations, supports anomaly analysis and root cause insight, outputs business control strategies, and finally realizes the closed-loop of operation management.

[0100] Figure 3 is the principle schematic diagram of the data analysis method provided by an embodiment of the present application. As Figure 3As shown, the application domain of the data analysis method according to the embodiments of the present application can be an inspector, an inventory checker, or a manager. Task scheduling can achieve intention recognition, meta-prompting, knowledge query, instruction execution, and result display (text, image). The model domain can include LLM models, embedding models, semantic segmentation models, and re-ranking models. The knowledge domain can include a text knowledge base composed of document loading interest rate increase, document segmentation, meta-information extraction, and vector representation, and a business database of the business system. It realizes data sharing and collaboration, effectively optimizes the business process, improves the data analysis efficiency and success rate, and improves the grass-roots management work efficiency.

[0101] Figure 9 is a schematic diagram of the main units of the data analysis device according to the embodiments of the present application. As Figure 9 shown, the data analysis device 900 includes an acquisition unit 901, a keyword determination unit 902, a query unit 903, and a data analysis unit 904.

[0102] The acquisition unit 901 is configured to, in response to a data analysis request, acquire the corresponding job role identifier and requirement text data;

[0103] The keyword determination unit 902 is configured to segment the requirement text data based on paragraph boundaries to obtain corresponding text blocks, input the text blocks into an embedding model to obtain corresponding text vectors, perform meta-information extraction on the text blocks to obtain corresponding meta-information, and input the text blocks into a large model to obtain corresponding keywords;

[0104] The query unit 903 is configured to call the domain knowledge base corresponding to the job role identifier to query and obtain corresponding prompt words based on the text vectors, meta-information, and keywords;

[0105] The data analysis unit 904 is configured to generate a data analysis result based on the prompt words.

[0106] In some embodiments, the data analysis device further includes Figure 9 a domain knowledge base construction unit not shown in the figure, which is configured to: acquire a list of job identifiers corresponding to a preset business, and acquire corresponding job knowledge data based on the list of job identifiers; convert the job knowledge data into text data, and then call a semantic segmentation model to perform text segmentation on the text data based on paragraph boundaries to obtain each text block; input each text block into an embedding model to obtain each text vector, perform meta-information extraction on each text block to obtain each meta-information, and input each text block into a large model to obtain corresponding keywords; acquire each risk response measure and corresponding prompt words corresponding to each job identifier in the list of job identifiers; construct a domain knowledge base based on each text vector, each meta-information, each keyword, each prompt word, and each risk response strategy.

[0107] In some embodiments, the data analysis device further includes Figure 9 an anomaly monitoring unit (not shown in the figure), which is configured to: obtain business operation data in real time, call an anomaly monitoring program to monitor in real time whether the business operation data is abnormal, and if so, trigger a data analysis request.

[0108] In some embodiments, the data analysis unit 904 is further configured to: obtain a risk response strategy corresponding to the prompt word from the domain knowledge base as the data analysis result.

[0109] In some embodiments, the data analysis unit 904 is further configured to: determine the corresponding process link according to the prompt word; generate the corresponding data analysis result based on the process link.

[0110] In some embodiments, the data analysis unit 904 is further configured to: in response to the process link being configured as model prediction, call the optimal hyperparameter combination model to perform data prediction based on the text vector, meta information, and keywords to obtain the predicted data analysis result.

[0111] It should be noted that the data analysis method and data analysis device of the present application have a corresponding relationship in the specific implementation content, so the repeated content will not be described again.

[0112] Figure 10 An exemplary system architecture 1000 to which the data analysis method or data analysis device of the embodiments of the present application can be applied is shown.

[0113] As Figure 10 shown, the system architecture 1000 may include terminal devices 1001, 1002, 1003, a network 1004, and a server 1005. The network 1004 is used to provide a medium for communication links between the terminal devices 1001, 1002, 1003 and the server 1005. The network 1004 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0114] Users can use the terminal devices 1001, 1002, 1003 to interact with the server 1005 through the network 1004 to receive or send messages, etc. Various communication client applications may be installed on the terminal devices 1001, 1002, 1003, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only as examples).

[0115] The terminal devices 1001, 1002, 1003 may be various electronic devices with a data analysis processing screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, and desktop computers, etc.

[0116] Server 1005 may be a server that provides various services. For example, it can be a background management server (only as an example) that supports data analysis requests submitted by users using terminal devices 1001, 1002, and 1003. The background management server can, in response to a data analysis request, obtain the corresponding job role identifier and requirement text data; split the requirement text data based on paragraph boundaries to obtain corresponding text blocks, input the text blocks into an embedding model to obtain corresponding text vectors, extract meta-information from the text blocks to obtain corresponding meta-information, input the text blocks into a large model to obtain corresponding keywords; call the domain knowledge base corresponding to the job role identifier to query and obtain corresponding prompt words based on the text vectors, meta-information, and keywords; and generate a data analysis result based on the prompt words. Thus, data sharing and collaboration can be achieved, the business process can be effectively optimized, the data analysis efficiency and success rate can be improved, and the grass-roots management work efficiency can be enhanced.

[0117] It should be noted that the data analysis method provided by the embodiments of the present application is generally executed by server 1005. Correspondingly, the data analysis device is generally set in server 1005.

[0118] It should be understood that Figure 10 the numbers of the terminal devices, networks, and servers in

[0119] are merely illustrative. According to the implementation requirements, there can be any number of terminal devices, networks, and servers. Figure 11 is a schematic structural diagram of a computer system 1100 of a terminal device suitable for implementing the embodiments of the present application. Figure 11 The terminal device shown is merely an example and should not impose any limitations on the functions and usage scope of the embodiments of the present application.

[0120] As Figure 11 shown, the computer system 1100 includes a central processing unit (CPU) 1101, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1102 or the program loaded from the storage section 1108 into the random access memory (RAM) 1103. In the RAM 1103, various programs and data required for the operation of the computer system 1100 are also stored. The CPU 1101, ROM 1102, and RAM 1103 are connected to each other through a bus 1104. The input / output (I / O) interface 1105 is also connected to the bus 1104.

[0121] The following components are connected to the I / O interface 1105: an input section 1106 including a keyboard, a mouse, etc.; an output section 1107 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc. and a speaker, etc.; a storage section 1108 including a hard disk, etc.; and a communication section 1109 including a network interface card such as a LAN card, a modem, etc. The communication section 1109 performs communication processing via a network such as the Internet. A drive 1110 is also connected to the I / O interface 1105 as required. A removable medium 1111, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is mounted on the drive 1110 as required so that a computer program read therefrom is installed into the storage section 1108 as required.

[0122] Specifically, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product that includes a computer program carried on a computer-readable medium, and the computer program includes program codes for performing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 1109, and / or installed from the removable medium 1111. When the computer program is executed by a central processing unit (CPU) 1101, the above functions defined in the system of the present application are executed.

[0123] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can, for example, include but is not limited to an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of a computer-readable storage medium can include but are not limited to: an electrical connection with one or more wires, 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), 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. In this application, a computer-readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, apparatus, or device. And in this application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, and this computer-readable medium can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on a computer-readable medium can be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0124] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram can represent a module, a program segment, or a part of code, and the above module, program segment, or part of code contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, as well as the combination of blocks in a block diagram or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0125] The units involved in the embodiments described in this application can be implemented in software or in hardware. The described units can also be provided in a processor. For example, it can be described as: a processor includes an acquisition unit, a keyword determination unit, a query unit, and a data analysis unit. Among them, the names of these units do not constitute a limitation on the units themselves in some cases.

[0126] As another aspect, the present application also provides a computer-readable medium. The computer-readable medium can be included in the device described in the above embodiments; it can also exist alone without being assembled into the device. The above computer-readable medium carries one or more programs. When the one or more programs are executed by the device, the device responds to a data analysis request, acquires the corresponding job role identifier and requirement text data; splits the requirement text data based on paragraph boundaries to obtain corresponding text blocks, inputs the text blocks into an embedding model to obtain corresponding text vectors, extracts meta-information from the text blocks to obtain corresponding meta-information, inputs the text blocks into a large model to obtain corresponding keywords; calls the domain knowledge base corresponding to the job role identifier to query corresponding prompt words based on the text vectors, meta-information, and keywords; and generates a data analysis result based on the prompt words.

[0127] According to the technical solution of the embodiments of the present application, data sharing and collaboration can be realized, the business process can be effectively optimized, the data analysis efficiency and success rate can be improved, and the grass-roots management work efficiency can be improved.

[0128] The above specific implementation manners do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present application shall be included within the protection scope of the present application.

Claims

1. A data analysis method, characterized in that, Including: In response to a data analysis request, obtain the corresponding job role identifier and requirement text data; Segment the requirement text data based on paragraph boundaries to obtain corresponding text blocks, input the text blocks into an embedding model to obtain corresponding text vectors, extract meta-information from the text blocks to obtain corresponding meta-information, and input the text blocks into a large model to obtain corresponding keywords; Call the domain knowledge base corresponding to the job role identifier to query for corresponding prompt words based on the text vectors, the meta-information, and the keywords; Generate a data analysis result based on the prompt words.

2. The method according to claim 1, wherein Before calling the domain knowledge base corresponding to the job role identifier, the method further includes: Obtain a list of job identifiers corresponding to a preset business, and obtain corresponding job knowledge data based on the list of job identifiers; Convert the job knowledge data into text data, and then call a semantic segmentation model to perform text segmentation on the text data based on paragraph boundaries to obtain each text block; Input each text block into an embedding model to obtain each text vector, extract meta-information from each text block to obtain each meta-information, and input each text block into a large model to obtain corresponding keywords for each; Obtain each risk response measure and corresponding prompt words corresponding to each job identifier in the list of job identifiers; Construct a domain knowledge base based on the text vectors, the meta-information, the keywords, the prompt words, and the risk response strategies.

3. The method according to claim 1, wherein Before obtaining the corresponding job role identifier and requirement text data, the method further includes: Obtain business operation data in real time, and call an anomaly monitoring program to monitor in real time whether the business operation data has anomalies. If so, trigger a data analysis request.

4. The method according to claim 1, characterized in that The generating a data analysis result based on the prompt words includes: Obtain a risk response strategy corresponding to the prompt word from the domain knowledge base as the data analysis result.

5. The method according to claim 1, characterized in that, The generating a data analysis result according to the prompt words includes: Determine a corresponding process link according to the prompt words; Generate a corresponding data analysis result based on the process link.

6. The method according to claim 5, characterized in that, The generating a data analysis result according to the prompt words includes: In response to the process link including model prediction, call an optimal hyperparameter combination model to perform data prediction based on the text vectors, the meta-information, and the keywords to obtain a predicted data analysis result.

7. A data analysis device, characterized in that, Including: An acquisition unit configured to obtain the corresponding job role identifier and requirement text data in response to a data analysis request; A keyword determination unit configured to segment the requirement text data based on paragraph boundaries to obtain corresponding text blocks, input the text blocks into an embedding model to obtain corresponding text vectors, extract meta-information from the text blocks to obtain corresponding meta-information, and input the text blocks into a large model to obtain corresponding keywords; A query unit configured to call the domain knowledge base corresponding to the job role identifier to query for corresponding prompt words based on the text vectors, the meta-information, and the keywords; A data analysis unit, configured to generate a data analysis result based on the prompt word.

8. The device according to claim 7, characterized in that The device further includes a domain knowledge base construction unit, configured to: Obtain a list of position identifiers corresponding to a preset service, and obtain corresponding position knowledge data based on the list of position identifiers; Convert the position knowledge data into text data, and then call a semantic segmentation model to perform text segmentation on the text data based on paragraph boundaries to obtain each text block; Input each of the text blocks into an embedding model to obtain each text vector, extract meta-information from each of the text blocks to obtain each meta-information, and input each of the text blocks into a large model to obtain corresponding keywords; Obtain various risk response measures and corresponding prompt words corresponding to each position identifier in the list of position identifiers; Construct a domain knowledge base based on the text vectors, the meta-information, the keywords, the prompt words, and the risk response strategies.

9. The device according to claim 8, characterized in that, The device further includes an anomaly monitoring unit, configured to: Obtain business operation data in real time, and call an anomaly monitoring program to monitor in real time whether the business operation data is abnormal. If so, trigger a data analysis request.

10. The device according to claim 8, characterized in that, The data analysis unit is further configured to: Obtain a risk response strategy corresponding to the prompt word from the domain knowledge base as a data analysis result.

11. A data analysis electronic device, characterized in that, Including: One or more processors; A storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1-6.

12. A computer-readable medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, the method according to any one of claims 1-6 is implemented.

Images