A Case-Based Reasoning with Artificial Intelligence Approach for Laptop Repair Assistant System
Keywords:
Case-Based Reasoning , Laptop Repair, Sentence-BERT, Semantic Similarity, Large Language Model (LLM)Abstract
This study presents a Case-Based Reasoning (CBR) approach for a Laptop Repair Assistant System, leveraging past repair cases to diagnose and resolve new technical issues efficiently. The system integrates Sentence-BERT (SBERT) for semantic feature extraction and LLaMA-3 (via Groq API) to dynamically revise solutions for partially matched cases. A dataset of 8,000 solved laptop issues was collected through web scraping, processed via text cleaning and SBERT embedding, and structured into a case base. The CBR cycle (Retrieve, Reuse, Revise, Retain) was implemented with cosine similarity thresholds: >90% for direct reuse, 60–90% for LLM-based revision, and <60% for user feedback requests. Evaluation through black-box testing confirmed system accuracy in exact matches, adaptive revisions, and handling novel cases. The system demonstrates scalability through automated case retention and user-friendly interfaces (Streamlit).
