基于LangChain与DeepSeek的JavaScript教学辅助智能体的研究

打开文本图片集

中图分类号：TP311；TP391.4 文献标识码：A 文章编号：2096-4706（2025）07-0076-07

Abstract： This paper proposes a localized inteligent teaching system based on LangChain and DeepSeek-7B-R1， aiming tosolve thechalengesofasynchronous programmingcomprehensionandDOMoperationerordetectionin JavaScript and jQuery teaching.Toaddresstheconflictsamong semanticunderstanding，privacysecurityandhardwarerequirements in traditional solutions，this paper designs a dual-engine architecture （rule pre-screening + LLMfine-tuning） thatintegratesdynamic ASTparsing andan enhancedRAGretrieval mechanism，constructinga teaching Knowledge Graphcovering 57typesof typical errors.Experiments demonstrate that thesystemachieves anaverage response time of 2.8 seconds on NVIDIA RTX 3060 device，and the teaching suggestion accuracy reaching 91 . 2 % ，witha 4 3 . 7 % improvement over conventionalLSTM approaches. Deployment in corporate internal training scenarios verifes its capabilityto enhance learner problem-solving eficiency by 41 % whilereducingteachermanual intervention by 76 % .Thisresearch provides the firsthigh-precisionand localized intelligent teaching framework supporting consumer-grade hardware for the programming education field， with open-source core components to lower development barriers.

Keywords： localized inteligent teachingsystem;LangChain rule engine; DeepSeek-7B-R1 fine-tuning; RAG mechanism; JavaScript Teaching Knowledge Graph;ASTdynamic parsing; consumer-grade hardware deployment; programming education efficiencyoptimization

0 引言

JavaScript作为Web开发的核心语言，占据GitHub项目总量的 2 8 . 3 % ，但其事件驱动、异步非阻塞等特性导致教学场景中普遍存在“概念理解-实践应用”断层现象。（剩余10126字）