Method, device, medium, and program product for generating an analysis model schema based on natural language input

Abstract

The application discloses a method, device, medium and program product for generating an analysis model Schema based on natural language input, the method comprising: using a deployed neural network model to perform intent recognition, business entity and analysis index extraction, time range analysis and output type judgment on analysis question text in natural language submitted by a user and organize the analysis question text into a semantic task structure; according to a predefined parameter mapping table, a domain index dictionary, a preconfigured analysis task type and analysis logic combination mapping table and an index and data source mapping relationship, writing corresponding parts in the semantic task structure into input parameters, field binding, analysis logic, data sources and output form fields in the analysis model Schema respectively, and generating a JSON analysis model Schema structure corresponding to the analysis question based on the fields. The application enables automatic and accurate modeling of natural language into an analysis model Schema structure with execution capability, and helps non-technical users to accumulate data modeling and organized data assets.

Description

Technical Field

[0001] This application relates to the field of electronic digital data processing, and in particular to a method for generating an analysis model schema based on natural language input. Background Technology

[0002] In data-driven decision-making scenarios, an increasing number of enterprises and organizations hope to quickly build data analysis models through natural language input to achieve operations such as attribution explanation of business phenomena, trend identification, and indicator monitoring. Currently, typical natural language analysis systems are mainly based on Text2SQL or NL2BI technologies, utilizing semantic parsing or large language model capabilities to directly convert user-submitted questions into structured query languages ​​or chart configuration requests, and then execute and return the results.

[0003] In practical applications, this type of technical solution has the following technical problems:

[0004] - Lack of intermediate semantic modeling mechanism between natural language and model structure. Current methods often directly convert natural language into SQL statements or chart configuration requests without structurally abstracting elements such as business intent, analysis dimensions, comparison logic, and anomaly detection methods, resulting in problems such as unstable semantic understanding, inaccurate field mapping, missing data conditions, and uncontrollable output structure;

[0005] - Lack of semantic structure separability and verifiability. Most of the conversion process is an uninterpretable one-time generation, and users cannot locate the source of errors in the semantic parsing process, nor can they correct or reconstruct parts of the structure, which limits the system's adaptability in multi-turn interactions and complex modeling;

[0006] - The system cannot form a standardized analytical model structure that is executable and reusable. Current systems typically only return data results and lack the ability to generate model schemas, making it impossible to save, share, or version the analysis process, and thus preventing it from being version-controlled and managed as an independent model asset.

[0007] - Low resource utilization and frequent repetitive calculations. The lack of caching mechanisms for structural and semantic results means that natural language needs to be re-parsed and query logic generated for each user query, which can easily lead to excessive system load and response delays in high-frequency interaction or large model forward call scenarios.

[0008] - Lack of model structural universality and platform decoupling capability. The model definition is strongly coupled with the platform and lacks portability to other data platforms and analysis platforms.

[0009] Therefore, there is a need to provide a method that can overcome at least some of the technical problems mentioned above. Summary of the Invention

[0010] This invention provides a method, device, medium, and program product for generating analytical model schemas based on natural language input. It can construct a complete semantic transformation path based on natural language input and automatically generate an executable, storable, and interpretable analytical model schema structure, which significantly improves the accuracy, usability, and engineering capabilities of intelligent modeling, thereby helping non-technical users with data modeling and organizations with data asset accumulation.

[0011] In a first aspect of the present invention, a method for generating an analysis model schema based on natural language input is provided, the method comprising the steps of:

[0012] S1: Receive analysis questions in natural language form submitted by users;

[0013] S2: Utilize the deployed neural network model to perform intent recognition, business entity and analysis indicator extraction, time range parsing, and output type judgment on the analysis question text. The recognized intent, extracted business entities, extracted analysis indicators, parsed time range, and judged output type are organized into a semantic task structure by corresponding to the analysis task type, analysis object, analysis indicator, time range, and expected output content in key-value pairs.

[0014] S3: Based on the predefined parameter mapping table, the analysis objects and time ranges in the semantic task structure are converted into the corresponding system modeling parameter formats, and the converted system modeling parameter formats are written into the input parameter fields in the analysis model schema;

[0015] S4: Based on the domain indicator dictionary, obtain the underlying field expression, unit, and data table source corresponding to the analysis indicator, and write the obtained underlying field expression into the field binding field in the analysis model schema;

[0016] S5: Based on the pre-configured mapping table between analysis task types and analysis logic combinations, select the corresponding analysis logic combination based on the analysis task type, and write the selected analysis logic combination into the analysis logic field in the analysis model schema.

[0017] S6: Based on the system's built-in mapping relationship between indicators and data sources, match the underlying data source for each analysis indicator and write the matched underlying data source into the data source field of the analysis model schema;

[0018] S7: Configure the output format according to the expected output content, and write the configured output format into the output format field in the analysis model schema;

[0019] S8: Generate a JSON analysis model schema structure corresponding to the analysis problem by combining the input parameters, field bindings, analysis logic, data source, and output format fields.

[0020] In a second aspect of the invention, a computer device is provided, including a processor, a memory, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in the first aspect.

[0021] In a third aspect of the invention, a computer-readable storage medium is provided having a computer program / instructions stored thereon, characterized in that the computer program / instructions, when executed by a processor, implement the steps of the method described in the first aspect.

[0022] In a fourth aspect of the invention, a computer program product is provided, comprising a computer program / instructions, characterized in that, when executed by a processor, the computer program / instructions implement the steps of the method described in the first aspect.

[0023] This invention achieves semantic isolation and logical decoupling between natural language and model structure by designing a unified semantic task structure object. It solves the semantic distortion and untraceable control problems caused by "direct generation of technical statements from natural language" in the existing Text2SQL solution, and helps to improve the system's controllability, interpretability and scalability.

[0024] By automatically mapping semantic task structures to model schema structures, a clear path for transformation from "business language" to "data modeling language" is established, improving the automation level of modeling, lowering the modeling threshold, and enabling non-technical users to participate in model generation and definition. During model schema generation, binding between metrics and fields automatically maps abstract business metrics to underlying data fields, ensuring semantic accuracy and structural stability of the model and improving its execution reliability. For different analysis task types, the system automatically configures and combines the logic layer by calling corresponding logic templates based on the semantic task structure, reducing the need for manual logic flow configuration.

[0025] In addition, the model schema constructed through this invention can be packaged into a standard model card, supporting saving, reuse, version control and cross-platform release, forming an organization-level manageable data modeling asset, breaking through the bottleneck of traditional one-time analysis of "instant query and disposable use".

[0026] Other features and advantages of the present invention will become clearer after reading the detailed description of the embodiments of the present invention in conjunction with the accompanying drawings. Attached Figure Description

[0027] Figure 1 This is a flowchart of an embodiment of the method according to the present invention.

[0028] For clarity, these figures are schematic and simplified, showing only the details necessary for understanding the invention, while omitting other details. Detailed Implementation

[0029] The embodiments and examples of the present invention will now be described in detail with reference to the accompanying drawings.

[0030] The scope of the invention will become apparent from the detailed description given below. However, it should be understood that while the detailed description and specific examples illustrate preferred embodiments of the invention, they are given for illustrative purposes only.

[0031] Figure 1 A flowchart illustrating a preferred embodiment of the method based on a natural language input generation analysis model schema according to the present invention is shown.

[0032] In step S1, the system receives analysis questions submitted by users in natural language form, such as: "Are the costs of cardiology departments abnormal in the past six months?" The submission of analysis questions can be done through web input boxes, voice recognition interfaces, BI platform integration portals, etc.

[0033] In step S2, the deployed neural network model is used to perform intent recognition, business entity and analysis indicator extraction, time range parsing, and output type determination on the analysis question text. The recognized intent, extracted business entities, extracted analysis indicators, parsed time range, and determined output type are organized into a semantic task structure in key-value pairs, corresponding to the analysis task type, analysis object, analysis indicator, time range, and expected output content, respectively.

[0034] For intent recognition, pre-trained neural network models (such as multi-classifiers based on LSTM or Transformer architectures) can automatically identify the analysis task type from the natural language questions input by users. For example, a trained intent classification model based on a Transformer encoder (such as a Large Language Model (LLM)) can be used to model the analysis question text, and then output the analysis task type label through a classification head. The training data for this intent classification model can be selected according to the application domain. For example, labeled medical operational analysis corpora can be used for supervised learning during the training phase. The task types that the model can recognize include: trend analysis, anomaly detection, comparison analysis, attribution analysis, and forecast analysis. For example, if the input question is "Has there been any abnormality in the workload of the cardiology department in the past six months?", the intent recognition result is "anomaly detection". "Anomaly detection" is mapped to "anomaly detection" and written into the analysis task type field (task_type) in the semantic task structure, forming the following structure:

[0035]

[0036] When the task type is unclear, the intent classification model automatically determines the closest question intent category (such as "trend analysis," "anomaly detection," or "comparative analysis") as the current analysis task type. This ensures that the model can still be built successfully even in the absence of explicit task prompts, demonstrating good fault tolerance and engineering stability.

[0037] For the extraction of business entities and analytical metrics, a trained named entity recognition method based on a Transformer encoder and pointer mechanism can be used to semantically parse the natural language analysis question input by the user, automatically extracting business entities (such as departments, units, and projects) and analytical metrics (such as costs, revenues, visits, and workloads). Specifically, the analysis question text is first segmented and part-of-speech tagged to form a token sequence. Then, a semantic encoding model based on the Transformer structure, commonly used in this field, is invoked to perform contextual modeling on the input token sequence through a multi-layer self-attention mechanism, obtaining a vector representation of the semantic position and semantic dependencies of each token (word) in the sentence. The semantic encoding model used can be, for example, a publicly available pre-trained model such as BERT, RoBERTa, or ELECTRA. In the entity recognition stage, a pointer mechanism is used to predict the start and end positions of each type of entity in the text and assign it a corresponding category label, such as "ORG_DEPT" (department) or "MEASURE" (metric). For example, when analyzing the question "Are the workload and costs of the cardiology department abnormal in the past six months?", the system can identify "cardiology department" as the business object and "workload" and "cost" as analysis indicators. After standardization mapping, the identification results are converted into a unified system encoding format, such as mapping "cardiology department" to "CARDIOLOGY" and "cost" to "cost_total", and written into the target_entity and metrics fields of the analysis object in the semantic task structure, forming the following structure:

[0038]

[0039] For time range parsing, trained neural network models can be used to automatically identify, structurally extract, and standardize time expressions contained in natural language text. For example, trained neural network models based on Transformer encoder structures such as BERT, RoBERTa, or ELECTRA, commonly used in this field, can be fine-tuned using corpora from application scenarios such as medical operations analysis. This allows for accurate identification and standardization of ambiguous, relative, and complex time expressions, recognizing time points or time intervals in natural language analysis text and automatically converting them into standard date format text that the system can process. For instance, using a large corpus of business queries containing time expressions as training data, this model can identify specific time points (such as "January 15, 2024") or time intervals (such as "last three months," "last quarter," "last year's data") in the text and automatically convert them into standard date format text that the system can process. For inputs with ambiguous time expressions, corresponding default values ​​can be provided as time ranges (such as "recent" automatically mapped to "last_3_months"). For example, if a user inputs, "Please analyze the changes in the average cost of surgical departments in the first half of the year," the system can recognize the time expression "first half of the year," parse it into a standard time range, standardize the recognized time range into a unified structure, and write it into the time range field of the semantic task structure, forming the following structure:

[0040]

[0041]

[0042] For output type determination, a model with the same structure as intent recognition can be used to identify the type of output the user desires. Multi-label output is supported, and recognizable output types include: charts, structured summaries, and yes / no answers (boolean answer). The identified output type is mapped to a system-unified encoding format and then written into the expected_output field of the semantic task structure, forming the following structure:

[0043]

[0044] In step S3, based on the predefined parameter mapping table, the analysis objects and time ranges in the semantic task structure are transformed into the corresponding system modeling parameter formats. For example, "cardiology" → "dept":"cardiology", "last six months" → "time_window":"last_6_months". The transformed system modeling parameter formats are then written into the input_params field of the analysis model schema, as shown in the example below:

[0045]

[0046] In step S4, based on the domain metric dictionary, the underlying field representation, unit, and data table source corresponding to the analysis metric are obtained. For example, "unit_cost" is bound to the field cost / workload and points to the fact_dept_costs table. The obtained underlying field representation is then written into the field binding metrics field in the analysis model schema, as shown in the following example:

[0047]

[0048] Each standard metric name in the domain metric dictionary includes its common synonyms, alternative names, and historically high-frequency expressions, enabling the matching of fuzzy user expressions (such as "spending") to standard metrics (such as "total_cost"). For example, the standard metric "total_cost" is associated with terms such as "spending," "expenses," and "total expenditure." When processing user input text, the system calculates the similarity between the extracted original terms and the metric dictionary, comprehensively using string edit distance, word vector semantic similarity (such as cosine similarity based on a pre-trained model), and rule engine judgment to determine the most likely corresponding standard metric. After a successful match, the system replaces the original expression with the standard metric name and writes it into the metrics field of the semantic task structure.

[0049] In step S5, based on the pre-configured mapping table between analysis task types and analysis logic combinations, the corresponding analysis logic combination is selected according to the analysis task type, and the selected analysis logic combination is written into the analysis logic logic field of the analysis model schema. For example, based on the analysis task type such as anomaly detection, the corresponding analysis template such as "trend change + threshold judgment" is selected and used as the configuration of the model logic layer.

[0050] During the deployment phase, a mapping table is pre-configured between "analysis task types" and "analysis logic combinations." Each task type corresponds to one or more standard logic module combinations. For example, the template for an anomaly detection task is "trend comparison + threshold judgment," and the template for the "comparative analysis" task is "group comparison + difference sorting." This mapping table is stored in the model schema in a structured configuration format and is defined as follows:

[0051] - Supported task types (e.g., anomaly_detection, trend_analysis, compare_analysis, etc.);

[0052] -Analysis logic components corresponding to each task type (e.g., trend_compare, deviation_threshold, dimension_breakdown, etc.);

[0053] - Parameter prerequisites required for specific logic (such as whether a time range or dimension field is required);

[0054] - Optional output suggestions (such as prioritizing the display of charts or conclusion text).

[0055] Example configuration is as follows:

[0056]

[0057] Once the system identifies the analysis task type (e.g., "anomaly_detection") from the semantic task structure, it performs the following processing:

[0058] -Based on the analysis task type, search for a matching logical combination in the analysis task type and analysis logic combination mapping table;

[0059] - Verify whether the semantic task structure contains input fields that meet the logical combination requirements for matching;

[0060] - If included, write the corresponding logical flow of the matching logical combination into the logic field of the analysis model schema;

[0061] - If not included, the default logical path will be written into the analysis logic field of the analysis model schema to generate basic statistical logic or simple comparative analysis logic.

[0062] The output example is as follows:

[0063]

[0064] For different analysis task types, the system calls corresponding logical templates based on semantic structure, achieving automatic configuration and combination of the logical layer, reducing the need for manual setting of logical processes. Simultaneously, the system supports automatic filling of parameters within the logical templates, such as determining whether multiple time windows are needed, whether grouping fields are included, and whether threshold judgment rules are introduced. The entire process requires no user intervention and features a configurable, scalable, and reusable template-based structure design. This significantly reduces the reliance on analytical experience in the modeling process, improves the automation and logical consistency of model building, and is particularly suitable for intelligent modeling applications aimed at non-technical users.

[0065] Automatic parameter filling for logical templates is completed during template instantiation. After identifying the task type (task_type), the system first matches the logical template library, and then automatically infers the required parameters based on the objects, metrics, and time range in the semantic task structure.

[0066] - Multiple time windows: Automatically determined by whether the time_range involves trend or comparative analysis;

[0067] - Grouping fields: Based on target_entity, map to corresponding fields such as department, doctor, or disease;

[0068] - Threshold rules: Read from the indicator threshold configuration table, such as unit cost deviation of 5%.

[0069] The inferred parameters will be written into the generated model schema, for example:

[0070] {

[0071] "logic":["trend_compare","deviation_threshold"],

[0072] "group_by":"dept_code",

[0073] "threshold":{"unit_cost":0.05}

[0074] }

[0075] The complete logical configuration can be guaranteed at the schema layer without manual intervention.

[0076] In step S6, based on the system's built-in mapping relationship between indicators and data sources, the underlying data source for each analysis indicator is matched, and the matched underlying data source is written into the data source field of the analysis model schema.

[0077] The system maintains a structured index data source mapping table to describe the data table name or API path that each index should use under different analysis dimensions. For example, "unit_cost" corresponds to the data table "fact_dept_costs" under the department dimension and "fact_doctor_costs" under the doctor dimension. Once the system identifies the analysis indices (such as "unit_cost" and "surgery_volume") and the analysis objects (such as "cardiology" and "Zhang San") in the semantic task structure, it determines the granularity (such as department, doctor, disease) based on the entity type and uses this as the basis to find the most matching data table or API path. Matching prioritizes dimension-related configurations; if no explicit dimension definition is provided, the default index data source is used as a fallback. The matching result is written to the data_source field in the model schema.

[0078]

[0079] This data source information will be used for data acquisition operations in the subsequent model scheduling and execution phases to ensure that the generated model is runnable and context-consistent.

[0080] In step S7, the output format, such as chart type (e.g., line chart, bar chart), text summarization rules, and structured JSON output template, is configured according to the expected output content in the semantic task structure. The configured output format is then written into the output field of the analysis model schema.

[0081]

[0082] The output format configuration is done automatically by the system. The expected output field, which the system identifies in the semantic task structure, is the result of the user's question being classified by the neural network intent recognition model. It includes the output type label that the user hopes to obtain, such as "chart", "structured_summary", "boolean_answer", etc.

[0083] When processing this field, the model schema generation module matches it against a predefined output template library. The output template library defines the specific representations for different output types, for example:

[0084] - When the recognition result contains "chart", the system will automatically recommend the most suitable chart type according to the task type, such as a line chart for trend analysis tasks, a bar chart for comparison analysis tasks, and a marked line chart for anomaly detection tasks.

[0085] - When the recognition result contains "structured_summary", the system will call the text summary template to generate a natural language description containing the main indicator values ​​and brief explanations;

[0086] - When the recognition result contains "boolean_answer", the system will configure a Boolean structured output (yes / no) and can attach a description of the triggering conditions.

[0087] For example, if the input question is "Has the workload of the cardiology department been abnormal in the past six months?", the expected_output in the semantic task structure is identified as ["chart","structured_summary"]. The system automatically generates the following schema fragment:

[0088] {

[0089] "output":["line_chart","structured_summary"]

[0090] }

[0091] "line_chart" comes from the default chart configuration when the task type is "anomaly detection", and "structured_summary" corresponds to the standard summary template.

[0092] In step S8, the input parameters, field bindings, analysis logic, data source, and output format fields are used to generate a JSON analysis model schema structure corresponding to the analysis question. The final generated schema structure is shown below:

[0093]

[0094] This schema structure can be used for model scheduling and execution, as well as as a carrier for model card persistence, version control, or deployment to the BI platform.

[0095] In an embodiment, the method of the present invention further includes validating the generated analysis model schema structure, detecting whether the fields are complete, whether the data source is available, and whether the logic matches the task type, thereby improving robustness and user experience under varying natural language input.

[0096] In this embodiment, the method of the present invention further includes encapsulating the generated analysis model schema structure into a JSON file and storing it in the analysis model library, assigning it a globally unique identifier and version number, and recording its status, creation time, modification time, etc. This transforms the model schema from "instant generation" to "manageable resource," enabling long-term accumulation and organizational-level application, reflecting the value of the technology platform and long-term business adaptability. It also supports canary releases, blue-green switching, and rollback operations at the version level, ensuring that new versions do not affect business continuity while being gradually verified. At the same time, it allows for full traceability of the model's status, compatibility, and historical records, ensuring the stability and controllability of the version evolution process.

[0097] This invention supports users describing their analysis needs in natural language. The system automatically parses the semantics, extracts key elements, and transforms them into a standardized modeling structure, ultimately forming an executable analysis model schema, achieving automatic modeling capability from natural language to model structure (NL→Schema). During system operation, users submit natural language analysis questions through the input interface, such as "Are the costs of cardiology departments abnormal in the past six months?" After receiving the question text, the system first performs semantic parsing, identifying the analysis task type (e.g., anomaly detection), analysis object (e.g., cardiology department), time range (e.g., the past six months), and analysis indicators (e.g., unit cost, workload, etc.), and determines the expected output type (e.g., conclusion or trend judgment). The parsing results are organized into structured semantic task objects to clarify the modeling logic and business semantic control. Subsequently, based on this structure, the system, according to internally defined mapping rules, indicator dictionaries, and data source configuration strategies, converts the semantic structure into an executable model schema, including components such as input parameters, data source, analysis logic, and output configuration. The generated model structure can be stored and supports version management, parameter reuse, and inter-platform calls.

[0098] In another embodiment, a computer device is provided, including a processor, a memory, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described above.

[0099] In another embodiment, a computer-readable storage medium is provided that stores a computer program / instructions thereon, characterized in that the computer program / instructions, when executed by a processor, implement the steps of the method described above.

[0100] In another embodiment, a computer program product is provided, including a computer program / instructions, characterized in that the computer program / instructions, when executed by a processor, implement the steps of the method described above.

[0101] The various embodiments described herein, or their specific features, structures, or characteristics, may be suitably combined in one or more embodiments of the invention. Furthermore, in some cases, the order of steps described in the flowcharts and / or pipeline processes may be modified where appropriate, and they need not be performed in the exact order described. Additionally, various aspects of the invention may be implemented using software, hardware, firmware, or combinations thereof, and / or other computer-implemented modules or devices that perform the described functions. Software implementations of the invention may include executable code stored in a computer-readable medium and executed by one or more processors. Computer-readable media may include computer hard disk drives, ROM, RAM, flash memory, portable computer storage media such as CD-ROM, DVD-ROM, flash drives, and / or other devices having a Universal Serial Bus (USB) interface, and / or any other suitable tangible or non-transitory computer-readable medium or computer memory on which executable code can be stored and executed by a processor. The invention may be used in conjunction with any suitable operating system.

[0102] Unless explicitly stated otherwise, the singular forms “a” and “the” used herein include the plural meaning (i.e., meaning “at least one”). It should be further understood that the terms “having,” “comprising,” and / or “including” as used in the specification indicate the presence of the described features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The term “and / or” as used herein includes any and all combinations of one or more of the listed related items.

[0103] The foregoing has described some preferred embodiments of the present invention. However, it should be emphasized that the present invention is not limited to these embodiments, but can be implemented in other ways within the scope of the present invention. Those skilled in the art can make various modifications and variations to the present invention based on the inventive concept and without departing from the scope of the present invention, and such modifications or variations still fall within the protection scope of the present invention.

Claims

1. A method for generating an analysis model Schema based on natural language input, characterized in that, The method includes the following steps: S1: Receive analysis questions in natural language form submitted by users; S2: Utilizing the deployed neural network model, the analysis question text is subjected to intent recognition, business entity and analysis indicator extraction, time range parsing, and output type determination. The recognized intent, extracted business entities, extracted analysis indicators, parsed time range, and determined output type are organized into a semantic task structure in key-value pairs, corresponding to the analysis task type, analysis object, analysis indicator, time range, and expected output content, respectively. The intent recognition includes: using a trained intent classification model based on a Transformer encoder to model the analysis question text, and then outputting analysis task type labels through a classification head, wherein the analysis task types include trend analysis, anomaly detection, comparative analysis, attribution analysis, and / or predictive analysis. S3: Based on the predefined parameter mapping table, the analysis objects and time ranges in the semantic task structure are converted into the corresponding system modeling parameter formats, and the converted system modeling parameter formats are written into the input parameter fields in the analysis model schema; S4: Based on the domain indicator dictionary, obtain the underlying field expression, unit, and data table source corresponding to the analysis indicator, and write the obtained underlying field expression into the field binding field in the analysis model schema; S5: Based on the pre-configured mapping table between analysis task types and analysis logic combinations, select the corresponding analysis logic combination based on the analysis task type, and write the selected analysis logic combination into the analysis logic field in the analysis model schema. S6: Based on the system's built-in mapping relationship between indicators and data sources, match the underlying data source for each analysis indicator and write the matched underlying data source into the data source field of the analysis model schema; S7: Configure the output format according to the expected output content, and write the configured output format into the output format field in the analysis model schema; S8: Generate a JSON analysis model schema structure corresponding to the analysis question by combining the input parameters, field bindings, analysis logic, data source, and output format fields; Step S5 includes: Based on the analysis task type in the semantic task structure, search for a matching logical combination in the pre-configured mapping table between analysis task type and analysis logical combination. Verify whether the semantic task structure contains input fields that meet the logical combination requirements for matching; If included, the corresponding logical flow of the matched logical combination will be written into the analysis logic field of the analysis model schema; If not included, the default logical path will be written to the analysis logical field of the analysis model schema.

2. The method of claim 1, wherein, The method further includes: The JSON analysis model schema structure is stored in the analysis model library and assigned a globally unique identifier and version number.

3. The method of claim 1, wherein, Each indicator in the domain indicator dictionary includes its standard name, synonyms, alternative names, and historically high-frequency expressions.

4. The method of claim 1, wherein, The extraction of business entities and analytical metrics includes: A trained named entity recognition method based on Transformer encoder and pointer mechanism is used to perform semantic parsing on natural language analysis questions input by users, and automatically extract business entities and analysis indicators.

5. The method of claim 1, wherein, The time range analysis includes: A trained neural network model based on the Transformer encoder structure is used to identify time points or time intervals in natural language analysis text and automatically convert them into standard date format text that the system can process.

6. A computer device comprising a processor, a memory and a computer program stored on the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1-5.

7. A computer readable storage medium having stored thereon computer programs / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method according to any one of claims 1-5.

8. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1-5.

Images