Abstract: The development of effective nucleic acid analysis such as gene chip technology in the medical field makes clinical diagnosis and medical research be able to apply it to acquire a large number of genetic information of tumor genesis. Meanwhile, with the progress and application of machine learning theory and technology recently, plenty of expert systems based on artificial intelligence technology have occurred in various fields. Aiming to the characteristics of the analyzed gene information, this paper proposes the methods and implementation of a tumor auxiliary diagnostic expert system, and discusses the key techniques of data mining on the system implementation process. It also describes the structural framework, working mechanism and auxiliary diagnosis principle of the system. The experiment releases the test results by using clinical medical data, and the result indicates that system implementation discussed herein can meet the requirement of an auxiliary diagnosis in certain extent.

Key words: genetic data mining, association rules, classification, auxiliary diagnosis, expert system

摘要： 在医疗领域中，基因芯片技术等高效核酸分析手段不断发展，使得临床诊断与医学研究中能够利用这一技术获取大量与肿瘤生成相关的基因信息。同时，近年来随着机器学习理论与技术的不断发展与应用，在各领域内出现了大量基于人工智能技术的专家系统。针对基因芯片信息的特点，描述了一种肿瘤辅助诊断专家系统的设计思路与实现方案；讨论了在专家系统实现过程中所采用的关键数据挖掘技术；重点叙述了系统的结构框架、工作机制与辅助诊断原理。在实验中，展示了临床获得的医疗数据在所论述系统中的测试结果。实验结果表明所论述的系统实现方案能够在一定程度上满足辅助诊断的需求。

关键词: 基因数据挖掘, 关联规则, 分类, 辅助诊断, 专家系统