Neural Network Software Requirement Effort Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current software development effort estimation processes are time-consuming and prone to human errors, leading to inaccurate and costly corrections of errors introduced early in the development phase.
Innovation Solution
A computer-implemented method using neural word vectors and deep learning to automatically evaluate software project requirements, estimating effort in measurable units and validating requirements for improved accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual effort estimation processes are used in software development, then human expertise and judgment can be applied to evaluate requirements, but the process becomes time-consuming and prone to human errors and bias
Solution Approach 1:
The patent replaces the manual mechanical process of effort estimation with an automated neural network-based system. The neural network engine processes requirement files and automatically generates effort estimates, validation scores, and story points without human intervention, thereby eliminating time consumption and human errors while maintaining or improving estimation accuracy.
Solution Approach 2:
The system enables self-service effort estimation by allowing the neural network to automatically evaluate requirements, perform validation, and generate estimates independently. The automated system serves itself by processing input requirement files and producing comprehensive output including effort estimates, validation scores, and confidence levels without requiring manual analysis.
2Reliability
If manual requirement evaluation is performed, then human reviewers can identify and correct errors, but errors introduced at early phases are often not detected until subsequent phases when correction costs multiply
Solution Approach 1:
The patent implements preliminary action by performing automated validation and error detection during the requirement analysis phase itself, before the development process begins. The neural network evaluates requirements for completeness, consistency, and clarity immediately, identifying potential errors early when correction is most efficient and cost-effective.
Solution Approach 2:
The system provides immediate feedback through validation scores and confidence levels that indicate the quality of requirement evaluation. This feedback mechanism allows for real-time identification and correction of issues, enabling continuous improvement and ensuring high reliability of the effort estimation process.
3Productivity
If automated effort estimation systems are implemented, then time consumption is reduced and consistency is improved, but the complexity of the system increases
Solution Approach 1:
The patent extracts the complex neural network processing logic into a separate, dedicated engine that operates independently. This modular approach allows the complex automated estimation system to function as a self-contained unit, reducing the apparent complexity for users while maintaining high productivity through automated processing.
Data Source
AI summary
Examples for automatically evaluating a software project requirement are disclosed. In an example, a neural word vector corresponding to a requirement file is generated and the neural word vector based on a score based vector is updated. An output vector comprising a conditional probability distribution of a plurality of answers associated with a plurality of questions identified from the updated neural word vector is generated. Further, a set of input parameters associated with at least one of the software project and the requirement is obtained. Based on the output vector and the set of input parameters, an effort required for completing the requirement may be estimated. A validation score associated with the requirement based on the output vector and a plurality of validation and classification parameters may be determined.