CN-121979493-A - Intelligent demand analysis and man-hour assessment method and system

CN121979493ACN 121979493 ACN121979493 ACN 121979493ACN-121979493-A

Abstract

The invention discloses a method and a system for intelligent demand analysis and man-hour assessment, which belong to the technical field of computer vision and artificial intelligence, wherein the method comprises the steps of constructing a multidimensional knowledge base; analyzing a multi-mode demand document containing texts, pictures and tables to generate structured demand data, carrying out module-level splitting on demands based on the structured demand data and system module description to generate a module task set, searching related system realization contexts for each task in the module task set based on a multi-dimensional knowledge base, fusing the structured demand data, the module task set and the system realization contexts to generate fine-grained technical tasks, calculating reference working hours for the fine-grained technical tasks based on historical case indexes, sequentially carrying out cross-module collaborative correction and personalized correction based on task executor capability on the reference working hours to obtain final evaluation working hours, and realizing automatic deep analysis and accurate and personalized working hour evaluation on software demands.

Inventors

DING ZHENHU
TIAN QIANG
MA GUANSHENG

Assignees

安徽中技国医医疗科技股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260116

Claims (12)

1. An intelligent demand analysis and man-hour assessment method, comprising: Constructing a multidimensional knowledge base which at least comprises a code index, a database index and a historical case index; analyzing the multi-mode demand document containing the text, the picture and the form to generate structured demand data; based on the structured demand data and the system module description, module-level splitting is carried out on the demand to generate a module task set; Retrieving an associated system implementation context for each task in the set of module tasks based on the multidimensional knowledge base; fusing the structured demand data, the module task set and the system realization context to generate fine-grained technical tasks; calculating reference man-hours for fine-grained technical tasks based on the historical case indexes; And sequentially performing cross-module collaborative correction and personalized correction based on the capability of the task executor on the reference working hours to obtain the final evaluation working hours.
2. The intelligent demand analysis and man-hour assessment method according to claim 1, wherein the parsing of the multimodal demand document comprises extracting text paragraphs in the document, identifying and extracting pictures embedded in the document, recording position marks thereof, identifying forms in the document and converting them into a structured data format.
3. The intelligent demand analysis and man-hour assessment method according to claim 2, wherein after identifying and extracting the picture embedded in the document, further comprising performing content analysis on the extracted picture through a visual model to identify flow chart nodes, interface layout or text content therein, and associating and integrating the analysis result with the text content.
4. The intelligent demand analysis and man-hour assessment method according to claim 1, wherein the module-level splitting of demands comprises determining software modules involved in demands based on the system module descriptions, and performing constraint correction on the split module task numbers by combining typical task numbers of similar demands in historical cases with module participation frequencies.
5. The intelligent demand analysis and man-hour assessment method according to claim 1, wherein the constructing a multidimensional knowledge base comprises constructing the code index by performing static analysis on source codes, wherein the code index comprises a back-end code index and a front-end page index, constructing the database index by reading database metadata, and constructing the historical case index by performing feature extraction and vectorization on historical worksheets.
6. The intelligent demand analysis and man-hour assessment method according to claim 5, wherein retrieving the associated system implementation context comprises selectively performing vector similarity retrieval in the back-end code index, front-end page index, database index and historical case index according to a current task type, and filtering and aggregating retrieval results based on module metadata.
7. The intelligent demand analysis and man-hour assessment method according to claim 1, wherein the calculating of the reference man-hour based on the history case index includes retrieving similar history cases in the history case index according to the description of the fine-grained technical task, obtaining an actual total man-hour of the similar history cases and a man-hour distribution ratio thereof on different task types, and distributing the reference total man-hour to the current fine-grained technical task according to the man-hour distribution ratio to form the reference man-hour.
8. The intelligent demand analysis and man-hour assessment method according to claim 1, wherein the cross-module collaborative correction is implemented by applying a non-linearly increasing adjustment factor as the number of system modules involved in the fine-grained technical task increases.
9. The intelligent demand analysis and man-hour assessment method according to claim 1, wherein the personalized correction based on the task performer's ability comprises determining a first adjustment factor according to the matching relation between the performer's job level and task complexity, determining a second adjustment factor according to the familiarity of the performer with the task-associated business module, and multiplying the reference man-hour by the first adjustment factor and the second adjustment factor to obtain corrected man-hour.
10. An intelligent demand analysis and man-hour assessment system, comprising: the multidimensional knowledge base module is used for storing and maintaining a code index, a database index and a historical case index; The document analysis module is configured to analyze the multi-mode requirement document to generate structural requirement data; The task disassembly engine is configured to generate a module task set based on the structured demand data and system module description, and search a system implementation context based on the multidimensional knowledge base module to generate fine-grained technical tasks; And the working hour evaluation module is configured to calculate the reference working hour based on the historical case index, and perform cross-module correction and personalized correction to obtain the final evaluation working hour.
11. The intelligent demand analysis and man-hour assessment system according to claim 10, wherein the multidimensional knowledge base module comprises: a code index unit for storing code semantic vectors extracted based on static analysis of source codes; A database index unit for storing a table structure and a relationship vector obtained based on reading database metadata; the historical case index unit is used for storing work order features and work hour distribution vectors obtained based on feature extraction and vectorization of historical work order data.
12. The intelligent demand analysis and man-hour assessment system according to claim 10 or 11, wherein the task disassembly engine comprises: A coarse splitting agent unit configured to perform the module level splitting; A fine-resolution proxy unit configured to retrieve a context based on the multidimensional knowledge base module and perform generation of the fine-grained technical task.

Description

Intelligent demand analysis and man-hour assessment method and system Technical Field The invention relates to the technical field of computer vision and artificial intelligence, in particular to an intelligent demand analysis and man-hour assessment method and system. Background In the software development life cycle, demand analysis and man-hour assessment are the core links of project planning stages, and accuracy of the software development life cycle directly influences the progress, cost and final delivery quality of a project. The need generally exists in the form of multimodal documents that contain not only a large number of textual descriptions, but also a rich picture, such as a business flow chart, a system architecture diagram, a user interface prototype diagram, etc., that bear vital logic and design requirements. Chinese patent publication No. CN118172031A discloses a project management man-hour assessment method, equipment and medium based on artificial intelligence, which are used for solving the problems of more data sources, complex data processing and lower accuracy and efficiency of the existing assessment method. The method comprises the steps of collecting multi-dimensional project data from a plurality of data sources of project management for marking, expanding new project sample data of the project management based on a SMOTE algorithm of distance measurement learning, carrying out feature extraction on the multi-dimensional project data and the new project sample data based on a feature extraction neural network, determining probability relations among features, carrying out feature dimension reduction on the project management based on a self-coding neural network according to the probability relations among the features, optimizing support vector selection of support vector machine algorithm based on decision boundary points to realize training of a classifier, classifying the multi-dimensional project data and the new project sample data after dimension reduction through the trained classifier, and evaluating project management man-hour according to the classified data. In the prior art, the resolving power of the picture information in the document is seriously insufficient, the processing steps in the flow chart and the interactive elements in the judgment logic or interface layout cannot be automatically identified and understood, so that a large amount of key demand information is lost or misunderstood in the analysis process, the demand analysis and task disassembly process is highly subjective, objective association with the existing system codes and database structures is lacking, the separated tasks possibly are disjointed with the actual technology, the cooperation complexity among the tasks and the influence of the individual capability difference of an executor on working hours cannot be effectively quantized, the fluctuation of the evaluation result is large, and the accuracy is low. Disclosure of Invention In order to solve the problems, the invention provides an intelligent demand analysis and man-hour assessment method and system, which adopts a means of constructing a multidimensional knowledge base, deeply analyzing the picture content in a demand document through a visual model, and carrying out context-related task disassembly and multidimensional man-hour correction by combining historical data, so that automatic deep analysis and accurate and personalized man-hour assessment on software demands can be realized. The above object can be achieved by the following scheme: An intelligent demand analysis and man-hour assessment method, comprising: Constructing a multidimensional knowledge base which at least comprises a code index, a database index and a historical case index; analyzing the multi-mode demand document containing the text, the picture and the form to generate structured demand data; based on the structured demand data and the system module description, module-level splitting is carried out on the demand to generate a module task set; Retrieving an associated system implementation context for each task in the set of module tasks based on the multidimensional knowledge base; fusing the structured demand data, the module task set and the system realization context to generate fine-grained technical tasks; calculating reference man-hours for fine-grained technical tasks based on the historical case indexes; And sequentially performing cross-module collaborative correction and personalized correction based on the capability of the task executor on the reference working hours to obtain the final evaluation working hours. Optionally, the parsing of the multimodal demand document includes extracting text paragraphs in the document, identifying and extracting pictures embedded in the document, recording position markers thereof, identifying forms in the document and converting them into a structured data format. Optionally, after identifying and extracting