Intelligent budgeting system for water conservancy informatization customization software development
By using an intelligent function point identification and analysis module, combined with a large language model and cost mapping knowledge base, the problems of imprecise function point division and insufficient compliance in the development of customized water conservancy information software have been solved. This has enabled rapid and accurate budget preparation and compliance assurance, reduced labor costs, and improved the system's operational efficiency and economic benefits.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- POWERCHINA BEIJING ENG CORP
- Filing Date
- 2026-01-30
- Publication Date
- 2026-05-12
AI Technical Summary
Conventional budgeting techniques suffer from several drawbacks in the development of customized water conservancy information software, including insufficient precision in functional point division, low efficiency, lack of real-time linkage capabilities, and inadequate compliance. These issues result in lengthy budgeting processes, inaccurate cost estimates, and high compliance risks.
The system employs an intelligent identification module, a multi-dimensional intelligent budget analysis module, and an intelligent report generation module. It combines an LLM large language model and a BERT model to identify and analyze functional points, constructs a water conservancy compliance cost mapping knowledge base, and realizes intelligent identification and compliance verification of functional points. Through parallel computing and cross-validation, it ensures that the identification accuracy rate reaches over 95%.
It enables rapid and efficient function point identification and dynamic and accurate pricing, reduces the cost of human involvement, improves budget response speed and compliance, ensures the legality and compliance of budget results, and enhances the system's operational efficiency and economic benefits.
Smart Images
Figure CN122022731A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information processing technology, and in particular to an intelligent budget preparation system for the development of customized software for water conservancy informatization. Background Technology
[0002] Conventional budgeting techniques still rely heavily on manual methods, which have significant limitations and struggle to meet the demands of accurate, predictable, and efficient budgeting for modern water resources information technology customized software development. Therefore, utilizing artificial intelligence to achieve intelligent transformation in budgeting for water resources information technology customized software development has become a crucial breakthrough for improving the efficiency of budgeting for water resources information technology projects and an important means to enhance budgeting efficiency.
[0003] Based on the analysis of the current state of the industry, the limitations of conventional budgeting techniques are mainly reflected in the following aspects:
[0004] (1) Low precision in functional point division leads to low compilation efficiency and long processing time: The conventional budget compilation technology for water conservancy informatization suffers from low efficiency and excessive processing time due to insufficient precision in functional point division. The project decomposition is too granular, resulting in incomplete cost coverage and difficulty in accurately reflecting the true costs of hardware, software, operation and maintenance, and integration. Minor adjustments to the project require repeated manual verification and adjustment, which not only increases the workload but also prolongs the approval and decision-making cycle, ultimately affecting the overall efficiency and timeliness of the compilation.
[0005] (2) Lack of real-time linkage capability: Conventional budget management methods are mostly one-time preparation and phased adjustment. When the project progress, design scheme or resource allocation changes, the budget needs to be recalculated and adjusted manually, which often leads to problems such as project delay or even budget overrun.
[0006] Therefore, there is an urgent need to provide an intelligent budget preparation system for customized software development of water conservancy information systems, which can improve timeliness and compliance compared to existing technologies. Summary of the Invention
[0007] The purpose of this invention is to provide: A smart budgeting system for water conservancy information customized software development, and related technologies, to solve technical problems such as how to improve timeliness and compliance, or a combination thereof.
[0008] Terminology Explanation: Unless otherwise defined, all technical terms used herein have the same meanings as commonly understood by one of ordinary skill in the art to which the subject matter pertains. Unless otherwise stated, all patents, patent inventions, and publications cited throughout this document are incorporated herein by reference in their entirety. Where multiple definitions exist for terms in this document, the definitions provided in this chapter shall prevail.
[0009] It should be understood that the above brief description and the following detailed description are exemplary and for illustrative purposes only, and do not limit the subject matter of the invention in any way. In this invention, the singular is used in conjunction with the plural unless otherwise specifically stated. It should also be noted that, unless otherwise stated, the use of “or” or “or” means “and / or”. Furthermore, the use of the term “comprising” and other forms such as “including,” “containing,” and “contains” are not limiting.
[0010] The definition of the standard terminology can be found in the reference "Benchmark Data of China's Software Industry in 2025".
[0011] Unless otherwise stated, conventional methods within the scope of the art, such as core budgeting methods, shall be used.
[0012] Unless otherwise defined, the use of various commercially available products as described herein employs standard techniques. These techniques and methods can generally be implemented according to conventional methods well-known in the art, based on the descriptions in the numerous general and more specific documents cited and discussed in this specification.
[0013] The terms “optional / arbitrary” or “optionally / arbitrarily” mean that the event or situation described below may or may not occur, including both the occurrence and non-occurrence of the event or situation.
[0014] The term "LLM Large Language Model" used in this article refers to a deep learning model that can understand and generate human language through training on massive amounts of text.
[0015] The term "BERT model" used in this article refers to a deep bidirectional Transformer encoder that understands text and adapts to various NLP tasks through two stages: pre-training and fine-tuning.
[0016] A smart budgeting system for water conservancy information customized software development includes a smart identification module, a multi-dimensional smart budget analysis module, and a smart report generation module. The smart identification module includes a data acquisition module and a smart classification module. The data acquisition module is used to acquire documents to be identified. The smart classification module is used to identify the functional point categories of the acquired documents. The multi-dimensional smart budget analysis module is used to obtain project costs based on the functional point categories identified by the smart classification module. The smart report generation module is used to display the project costs obtained by the multi-dimensional smart budget analysis module.
[0017] Furthermore, the intelligent classification module includes a water conservancy compliance cost mapping knowledge base, and also trains an LLM large language model and a BERT model for function point category analysis.
[0018] Furthermore, in the intelligent classification module, the documents to be identified are divided into two categories based on document type: structured short text and unstructured long text. The trained LLM large language model is used to extract function points from the unstructured long text, and the trained BERT model is used to extract function points from the structured short text. When the function points extracted by the two models are water conservancy professional function points, the function point category is output to the multi-dimensional intelligent budget analysis module; otherwise, it is identified as a regular function point, and the category recognition is verified for the regular function points.
[0019] Furthermore, the method for verifying the category identification of common functional points is as follows: obtain the confidence level of the BERT model or LLM large language model when identifying the category of common functional points. When the confidence level of the BERT model or LLM large language model after identifying the type of common functional point is greater than or equal to 90%, the category of the common functional point is output to the multi-dimensional intelligent budget analysis module; if the confidence level is less than 90%, the pre-rule base verification is triggered.
[0020] Furthermore, the specific method for pre-validation of the rule base is as follows: The category of the regular function point is matched against the water conservancy compliance cost mapping knowledge base. If the function point exists in the water conservancy compliance cost mapping knowledge base, its category is output to the multi-dimensional intelligent budget analysis module. If the function point does not exist in the water conservancy compliance cost mapping knowledge base, it is input into the lightweight classifier to obtain the confidence level of the lightweight classifier's category identification for the function point. When the confidence level of the lightweight classifier's category identification for the function point is greater than or equal to 85%, the category identified by the lightweight classifier for the function point is output. When the confidence level is less than 85%, the function point category is input into the LLM large language model for recognition and outputs the function point category after recognition by the LLM large language model. The function point categories recognized by the lightweight classifier and the function point categories recognized by the LLM large language model are both triggered to verify the rule base. The function point category is matched with the water conservancy compliance cost mapping knowledge base. If they match, the function point category is input into the multi-dimensional intelligent budget analysis module. If they do not match, they are manually reviewed and then input into the multi-dimensional intelligent budget analysis module.
[0021] Furthermore, the multi-dimensional intelligent budget analysis module includes a data access module, a classification calculation module, and a compliance verification module. The data access module is used to obtain the function point categories output by the intelligent classification module. The classification calculation module is used to calculate project costs based on the function point categories. The compliance verification module is used to verify whether the calculated project costs are compliant.
[0022] Furthermore, project costs are calculated using the following formula: Project Costs = Project Workload × Person-Month Rate × Direct Non-Labor Costs.
[0023] Furthermore, the project workload is specifically calculated using the following formula: Project workload = (conventional AFP / conventional productivity) + (hydraulic AFP / hydraulic productivity); Conventional productivity = Conventional AFP / (Conventional productivity + Water conservancy AFP); Water productivity = Water AFP / (Conventional productivity + Water AFP).
[0024] Furthermore, conventional AFP and water conservancy AFP are specifically calculated using the following formula: Standard AFP = UFP × Size Change Adjustment Factor × Software Category Adjustment Factor × Software Quality Characteristics Adjustment Factor × Domestic Information Technology Innovation Adjustment Coefficient; Water conservancy AFP = UFP × CF; Where UFP represents the number of function points that have not been adjusted, and CF represents the set size change factor.
[0025] Furthermore, UFP is specifically calculated using the following formula: UFP= ; In the above formula, This represents the weight of each category. This indicates the total number of categories identified by the intelligent classification module.
[0026] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This invention intelligently identifies functional points, responds to market price fluctuations in real time, and improves work efficiency. It employs a dual-model collaborative verification mechanism, parallel computing, and cross-validation to ensure that the accuracy of functional point identification reaches over 95%, achieving rapid and efficient intelligent classification. Its dynamic and accurate pricing capability responds to market price fluctuations in real time, significantly improving budget response speed and adjustment flexibility, and providing favorable data support for early-stage project decision-making.
[0027] (2) This invention reduces compliance risks and improves compliance: full-process compliance assurance reduces policy risks. The system constructs a compliance assurance mechanism covering the entire budget preparation process, ensuring that budget results are legal and compliant, and risks are controllable, effectively avoiding compliance issues caused by policy misunderstandings or updates.
[0028] (3) This invention reduces the cost of human involvement: It reduces overall operating costs and improves long-term economic benefits. The intelligent budget preparation system is effective in reducing reliance on human resources and optimizing resource allocation, significantly reducing the frequency and workload of human involvement, reducing the cost of training professional personnel, thereby improving the system's sustainable operation and long-term economic benefits. Attached Figure Description
[0029] Figure 1This is a structural diagram of the system of the present invention. Detailed Implementation
[0030] The technical solution of the present invention will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention. All other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.
[0031] like Figure 1 As shown, this invention provides an intelligent budget preparation system for customized software development of water conservancy information systems, including an intelligent identification module, a multi-dimensional intelligent budget analysis module, and an intelligent report generation module; the intelligent identification module includes a data acquisition module and an intelligent classification module; the data acquisition module is used to acquire documents to be identified, and the intelligent classification module is used to identify the functional categories of the acquired documents to be identified.
[0032] The intelligent classification module includes a knowledge base for mapping water conservancy compliance costs. This knowledge base comprises a government price database and an industry database. The industry database fields include software development productivity for various industries, software development productivity for various regions, benchmark monthly software development costs, and specific values for the various factors involved in this invention. The intelligent classification module trains an LLM (Large Language Model) and a BERT (Browser Expert Model) to perform intelligent functional point analysis. Based on document type, documents to be identified are divided into two categories: structured short text and unstructured long text. Unstructured long text and structured short text are distinguished based on text length, formatting regularity, and whether semantic parsing is required. Text without a fixed format, longer than a length threshold, and requiring semantic parsing is classified as unstructured long text; text with a fixed format, shorter than or equal to the length threshold, and not requiring semantic parsing is classified as structured short text.
[0033] In the intelligent classification module, function points are extracted from unstructured long texts using a trained LLM large language model, and from structured texts using a BERT model. The module determines whether the function points extracted by both models contain water conservancy features. If they do, they are marked as water conservancy professional function points, and the function point category is output to the multi-dimensional intelligent budget analysis module. If they do not contain water conservancy features, they are marked as general function points. The module then identifies the category of each general function point, obtaining the confidence score of the BERT or LLM model in identifying that category. When the confidence score of the BERT or LLM model for identifying a general function point type is greater than or equal to 90%, the category of that general function point is output to the multi-dimensional intelligent budget analysis module. If the confidence score is less than 90%, a pre-defined rule base verification is triggered. The specific method for pre-validation of the rule base is as follows: The category of the regular function point is matched against the water conservancy compliance cost mapping knowledge base. If the function point exists in the knowledge base, its category is output to the multi-dimensional intelligent budget analysis module. If the function point does not exist in the knowledge base, it is input into a lightweight classifier to obtain the confidence level of the classifier's identification of the function point's category. When the confidence level of the lightweight classifier's identification of the function point's category is greater than or equal to 85%, the category identified by the lightweight classifier is output. When the confidence level is less than 85%, it is input into the LLM large language model for identification, and the function point category identified by the LLM large language model is output. Whether the function point category is identified by a lightweight classifier or by an LLM large language model, a rule base for triggering verification is applied. The function point category is matched with the water conservancy compliance cost mapping knowledge base. If a match is found, the function point category is input into the multi-dimensional intelligent budget analysis module. If a match is not found, manual review is performed. After manual review, the function point category is input into the multi-dimensional intelligent budget analysis module.
[0034] The characteristics of water conservancy include business scenario characteristics, professional terminology characteristics, and technical function characteristics. Among them, business scenario characteristics include flood early warning and simulation, reservoir scheduling and management, water resource matching and monitoring, and safety monitoring of water conservancy projects; professional terminology characteristics include flood, runoff, water level, flow, reservoir, gate, pumping station, embankment, water resources, water quality, water supply, water conservation, flood control and drought relief, soil erosion, and ecological flow; and technical function characteristics include hydrological model calculation, three-dimensional visualization of water conservancy projects, multi-source data fusion analysis, and real-time monitoring and early warning push.
[0035] The multi-dimensional intelligent budget analysis module includes a data access module, a classification calculation module, and a compliance verification module. The data access module is used to access external data sources such as the water conservancy project cost information network in real time and the categories of functional points output by the intelligent identification module, to obtain the latest labor cost rates and industry productivity standards, and to establish a dynamic data update mechanism to ensure the timeliness and accuracy of the pricing basis.
[0036] The classification calculation module is used to calculate project costs based on the data accessed by the data access module, specifically calculated using the following formula: Project cost = Project workload × Person-month rate × Direct non-human costs.
[0037] Project workload = (conventional AFP / conventional productivity) + (hydraulic AFP / hydraulic productivity).
[0038] Conventional productivity = Conventional AFP / (Conventional productivity + Water conservancy AFP).
[0039] Water productivity = Water AFP / (Conventional productivity + Water AFP).
[0040] Standard AFP = UFP × Size Change Adjustment Factor × Software Category Adjustment Factor × Software Quality Characteristics Adjustment Factor × Information Technology Innovation Adjustment Coefficient.
[0041] Water conservancy AFP = UFP × CF.
[0042] UFP= .
[0043] Where UFP represents the number of function points not adjusted, and CF represents the set size change factor. This represents the weight of each category. This indicates the total number of categories identified by the intelligent classification module.
[0044] The monthly labor cost rate, direct non-human costs, scale change adjustment factor, software category adjustment factor, software quality characteristic adjustment factor, and information technology innovation adjustment coefficient were all obtained from the "2025 China Software Industry Benchmark Data".
[0045] The compliance verification module is used to verify project costs against water conservancy informatization policies and regulations to ensure that the composition of each cost complies with relevant regulations. If the result does not comply with the regulations, the system will trigger an early warning mechanism for review, and the classification calculation module will recalculate the project costs until they comply with the regulations.
[0046] The intelligent report generation module includes a budget detail table module and a budget preparation report module. The budget detail table module generates a detailed budget table for the development of customized water conservancy information software based on the budget results, clearly listing the name, count, unit price, and total price of various functional points. The budget preparation report module generates a budget preparation report for the development of customized water conservancy information software based on the budget detail table, including complete content such as preparation instructions, a list of functional points, detailed cost calculations, and a compliance statement, providing a comprehensive basis for project decision-making and subsequent auditing.
[0047] This invention intelligently identifies functional points, responding in real-time to market price fluctuations and improving work efficiency. Employing a dual-model collaborative verification mechanism, parallel computing, and cross-validation, it ensures a functional point identification accuracy rate of over 95%, achieving rapid and efficient intelligent classification. Its dynamic and accurate pricing capabilities, responding to market price fluctuations in real time, significantly improve budget response speed and adjustment flexibility, providing valuable data support for early-stage project decision-making.
[0048] This invention reduces compliance risks and improves compliance: It provides end-to-end compliance assurance and reduces policy risks. The system establishes a compliance assurance mechanism covering the entire budget preparation process, ensuring that budget outcomes are legal, compliant, and risk-controllable, effectively avoiding compliance issues caused by misunderstandings or updates to policies.
[0049] This invention reduces the cost of human involvement: lowering overall operating costs and improving long-term economic benefits. The intelligent budgeting system is highly effective in reducing reliance on human resources and optimizing resource allocation, significantly reducing the frequency and workload of manual intervention, lowering the cost of training professional personnel, thereby enhancing the system's sustainable operation and long-term economic benefits.
[0050] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, and is not intended to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions made by those skilled in the art to the technical solution of the present invention do not depart from the essence and scope of the technical solution of the present invention.
Claims
1. An intelligent budgeting system developed using customized software for water conservancy information systems, characterized in that: The system includes an intelligent identification module, a multi-dimensional intelligent budget analysis module, and an intelligent report generation module. The intelligent identification module comprises a data acquisition module and an intelligent classification module. The data acquisition module is used to acquire documents to be identified. The intelligent classification module is used to identify the functional point categories of the acquired documents. The multi-dimensional intelligent budget analysis module is used to obtain project costs based on the functional point categories identified by the intelligent classification module. The intelligent report generation module is used to display the project costs obtained by the multi-dimensional intelligent budget analysis module.
2. The intelligent budgeting system for customized water conservancy information technology development according to claim 1, characterized in that, The intelligent classification module includes a water conservancy compliance cost mapping knowledge base. It also trains an LLM large language model and a BERT model for function point category analysis.
3. The intelligent budget preparation system for customized water conservancy information software development according to claim 2, characterized in that, In the intelligent classification module, the documents to be identified are divided into two categories according to document type: structured short text and unstructured long text. The trained LLM large language model is used to extract function points for unstructured long text, and the trained BERT model is used to extract function points for structured short text. When the function points extracted by the two models are water conservancy professional function points, the function point category is output to the multi-dimensional intelligent budget analysis module; otherwise, it is identified as a regular function point, and the category recognition is verified for regular function points.
4. The intelligent budget preparation system for customized water conservancy information technology software development according to claim 3, characterized in that, The verification method for class identification of common function points is as follows: obtain the confidence level of the BERT model or LLM large language model when identifying the class of common function points. When the confidence level of the BERT model or LLM large language model after identifying the type of common function point is greater than or equal to 90%, the class of the common function point is output to the multi-dimensional intelligent budget analysis module; if the confidence level is less than 90%, the pre-rule base verification is triggered.
5. The intelligent budgeting system for customized water conservancy information technology development according to claim 4, characterized in that, The specific method for pre-validation of the rule base is as follows: match the category of the regular function point with the water conservancy compliance cost mapping knowledge base. If the function point exists in the water conservancy compliance cost mapping knowledge base, output the category of the function point to the multi-dimensional intelligent budget analysis module; if the function point does not exist in the water conservancy compliance cost mapping knowledge base, input it into the lightweight classifier and obtain the confidence level of the lightweight classifier after classifying the category of the function point; when the confidence level of the lightweight classifier after classifying the category of the function point is greater than or equal to 85%, output the category identified by the lightweight classifier for the function point. When the confidence level is less than 85%, the data is input into the LLM large language model for recognition, and the function point category after recognition by the LLM large language model is output. The function point categories recognized by the lightweight classifier and the function point categories recognized by the LLM large language model are both triggered to verify the rule base. The function point categories are matched with the water conservancy compliance cost mapping knowledge base. If they match, the function point category is input into the multi-dimensional intelligent budget analysis module. If they do not match, they are manually reviewed. After manual review, the function point category is input into the multi-dimensional intelligent budget analysis module.
6. The intelligent budgeting system for customized water conservancy information technology development according to claim 1, characterized in that, The multi-dimensional intelligent budget analysis module includes a data access module, a classification calculation module, and a compliance verification module. The data access module is used to obtain the function point categories output by the intelligent classification module. The classification calculation module is used to calculate project costs based on the function point categories. The compliance verification module is used to verify whether the calculated project costs are compliant.
7. The intelligent budgeting system for customized water conservancy information technology development according to claim 6, characterized in that, Project costs are calculated using the following formula: Project Costs = Project Workload × Monthly Rate × Direct Non-Labor Costs 8. The intelligent budget preparation system for customized water conservancy information software development according to claim 7, characterized in that, The project workload is calculated using the following formula: Project workload = (conventional AFP / conventional productivity) + (hydraulic AFP / hydraulic productivity); Conventional productivity = Conventional AFP / (Conventional productivity + Water conservancy AFP); Water productivity = Water AFP / (Conventional productivity + Water AFP).
9. The intelligent budget preparation system for customized water conservancy information technology software development according to claim 8, characterized in that, Conventional AFP and water conservancy AFP are specifically calculated using the following formula: Standard AFP = UFP × Size Change Adjustment Factor × Software Category Adjustment Factor × Software Quality Characteristics Adjustment Factor × Domestic Information Technology Innovation Adjustment Coefficient; Water conservancy AFP = UFP × CF; Where UFP represents the number of function points that have not been adjusted, and CF represents the set size change factor.
10. The intelligent budgeting system for customized water conservancy information technology development according to claim 9, characterized in that, UFP is specifically calculated using the following formula: UFP= ; In the above formula, This represents the weight of each category. This indicates the total number of categories identified by the intelligent classification module.