Software Bug Knowledge Graph Construction via NLP
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current software bug search methods based on traditional relationship databases are inefficient due to inaccurate and irrelevant search results, making it difficult for developers and maintainers to find necessary information quickly and effectively.
Innovation Solution
An automatic construction method for a software bug knowledge graph using natural language processing to extract relationship triples from bug descriptions, combined with semi-supervised support vector machine classification, to improve bug classification and information retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional relationship database search method is used, then bug information can be stored and retrieved, but search accuracy and relevance deteriorate
Solution Approach 1:
The patent replaces the traditional mechanical keyword-matching database search system with a knowledge graph-based semantic search system. The knowledge graph uses natural language processing to extract semantic relationships from bug descriptions, enabling the system to understand and retrieve information based on meaning rather than just keyword matching, thereby significantly improving search accuracy and result relevance.
Solution Approach 2:
The patent introduces a knowledge graph as an intermediary layer between the raw bug data and the search query. This intermediary structure processes and organizes bug information through natural language processing, extracting entities and relationships to form a semantic representation that mediates between user search intent and stored bug data, enabling more accurate and relevant search results.
2Productivity
If keyword matching search is used, then search can be performed, but time consumption increases due to manual browsing and filtering
Solution Approach 1:
The patent replaces manual keyword matching and browsing with an automated knowledge graph search system. The system automatically extracts semantic relationships from bug descriptions and organizes them in a knowledge graph structure, enabling automated retrieval and ranking of relevant bugs based on semantic similarity rather than manual filtering, thus significantly improving search efficiency and reducing time loss.
Solution Approach 2:
The knowledge graph system performs self-service by automatically organizing, indexing, and ranking bug information based on semantic relationships extracted through natural language processing. The system autonomously processes search queries by matching semantic meanings rather than requiring manual intervention for browsing and filtering, thereby improving productivity and reducing time consumption.
3Measurement precision
If natural language processing is applied to extract relationship triples, then bug information accuracy improves, but processing complexity increases
Solution Approach 1:
The patent segments the complex natural language processing task into distinct modular components: word segmentation, part-of-speech tagging, noun phrase extraction, verb phrase extraction, and relationship triple construction. Each component handles a specific aspect of the processing pipeline, making the overall complex task manageable and easier to implement while maintaining high accuracy in extracting bug information relationships.
Data Source
AI summary
Provided is an automatic construction method of a software bug knowledge graph. The method includes extraction of a relationship triple of a bug and domain classification of the bug. Specifically, the method includes: collecting bug information in a bug library and processing bug description information, obtaining a verb phrase and a noun phrase in a description sentence by means of natural language processing, and then obtaining a relationship triple of the bug according to a dependency relationship between words related to the bug information, extracting a domain feature of the bug, performing learning and training with a semi-supervised classifier to enable the classifier automatically to classify unlabeled triples, storing all the classified relationship triples in a graph database, and thus constructing a software bug knowledge graph.