台灣指數股票型基金（Exchange-traded funds,ETF）市場於台灣蓬勃發展，滿足各類投資人多元化之資產配置，只需於集中市場買進ETF，便可達到買賣標的指數一籃子股票之效果，而台灣50為台灣第一支交易型證券指數，對於操作選股時，亦為重要之指標，其成份股涵蓋台灣證券市場市值前50大上市公司，於各產業中處重要之支配地位，具營收穩定、交易量活躍及市場認同度高之特點。本研究運用資料探勘中關聯性法則技術（Association rules），選擇關聯法則中的Apriori演算法，以支持度、可靠度與增益值為門檻，尋找有用之規則，探討台灣集中市場各上市類股、財團法人中華民國證券暨期貨市場發展基金會第11屆資訊揭露評鑑A++等級股票、外資券商評等薦買股及台灣50成份股之關聯性，再以資料探勘技術找尋關聯性規則，於台灣指數股票型基金（ETF）發展建議，進而提供投資組合之概念及建議，作為投資人投資決策之參考。 Exchange-traded funds（ETF）market flourishes in Taiwan’s stock exchange market. It satisfies the demand of diversified asset allotments from investors. When ETF is bought, it has the effect that a basket of stocks are invested. As Taiwan’s first ETF, Taiwan 50 provides a major benchmark for selecting stocks on market. Since it comprises 50 largest listed companies, each dominating the market it is in, Taiwan 50 is characterized by stable revenues and sprightly transactions, and its component companies are identified as prestigious enterprises. This thesis uses data mining approach, association rules, and implements Apriori algorithm to investigate data mining results. By doing so, various Taiwan Stock Exchange listed stocks, stocks ranking A++ in transparency and disclosure by Securities & Futures Institute, and possible stocks selection recommended by foreign brokerage firms are proposed alternative components on Taiwan 50. In addition, possible portfolio suggestions from research findings are further discussed on this thesis.
