以FCM為基礎之演算法於微陣列血癌資料探勘之研究

基因微陣列分析速度快且專一性高，其主要是用來觀測基因的差異及變化。近年來研究發現單一個體的基因突變與疾病發生有關聯，而我們如何從龐大的基因序列中尋找到關鍵的基因組合變得十分重要。癌症於國人十大死因中蟬聯榜首多年，如何降低、預防癌症的發生一直是醫療領域的重點。本研究主要目的是找尋出最佳基因組合及基因彼此間關聯性，透過建構一套混合進化式演算法對基因微陣列血癌資料進行探勘，找尋關鍵病變屬性集合以判別癌症之高準確度的方法。本研究欲建立最佳之模糊認知圖進行癌症類別的分類，然而於文獻得知過去的研究中，模糊認知圖的變數、權重大都藉由專家經驗及建議所擬定，難以用客觀角度對變數組合進行選取，所以本研究跳脫傳統模糊認知圖的做法，發展出一套不需專家干預與經驗知識評量的研究方法，並透過人工智慧演算法達到學習、演化的自我學習演算法。由本研究實驗數據證實所提方法皆優於或等同於過去文獻的方法，顯示所提方法之優越性。

關鍵字

模糊認知圖；血癌；基因微陣列

並列摘要

DNA microarray can be used to analyze the specified data efficiently and effectively so as provide the observations of gene expression differences and their changes among genes. In recent years, some studies in literature have pointed out that mutation of a single gene often associated with concurrent disease so that how to find the key of genes combination from DNA sequence becomes a very important issue. Cancers are still the top ten causes of death in past decades, how to diagnosis the cancer in very early occurrence has been a major plan of the medical field. The aim of this study is to investigate how to find the best genes combination and relationship strength for identifying the categories of cancers from the microarray data. In this paper, a hybrid meta-evolutionary approach with grid computing architecture to assess microarray data pattern in the cancer classification problems is proposed for extracting the fuzzy cognitive map including the predictors, the corresponding parameters of the map simultaneously so as to building a decision making model with maximum classification accuracy. Without experts’ attendance to construct the fuzzy cognitive map, it can be developed in the proposed approach based on the numerical data to do decision analytic, it is bound to bring a new breakthrough in FCM applications. Through the numerical experiments, we compared our results against the methods in literature and the commercial data mining software, and then we show experimentally that the proposed approach is promising for improving prediction accuracy and enhancing the modeling simplicity.

並列關鍵字

Fuzzy Cognitive Maps ； Leukemia ； DNA microarray

參考文獻

[1]Zhigang Liana, Bin Jiaoa, Xingsheng Gu, 2006, “A similar particle swarm optimization algorithm for job-shop scheduling to minimize makespan”, Applied Mathematics and Computation, 183(2), pp.1008-1017.

Google Scholar

[2]Jihoon Yang, Rajesh Parekh, Vasant Honavar, 1999, “DistAl: An inter-pattern distance-based constructive learning algorithm”, Intelligent Data Analysis, 3(1), pp.55-73.