半導體產業是台灣發展的重點產業主要包含了IC設計、製造、封裝、測試,隨著資訊科技進步、產品的彈性提升、要求客製化的生產型態增加,因半導體產業的供應鏈複雜且競爭激烈,因此若想要在半導體產業投資有顯著的效果,必須對其產業的各個公司各種屬性進行有效的分析。本研究著眼於線性判別分析和決策樹兩種分析器建立半導體的投資決策支援系統,本研究擬以104年12月到105年12月以50家為訓練樣本,並選擇50家同性質產業的公司進行驗證,(a)以原始資料所得的正確率和投資報酬率(b)以決策樹和線性判別分析篩選所得的正確率(c)以決策樹和線性判別分析篩選所得的投資報酬率。
The semiconductor industry is a major industry in the development of Taiwan. It mainly includes IC design, manufacturing, IC packaging, and IC testing. With the improvement of information technology, the enhancement of product flexibility, and the demand for customized production. In the meantime, the supply chain of the semiconductor industry is complex and competitive. Therefore, the investment considering in the semiconductor industry has a significant effect, an effective analysis with employing various attributes of each company in the industry. This study focuses on linear discriminant analysis and decision tree to establish a semiconductor investment model on a decision support system. This topic is planned to use 50 training samples from December 104 to December 105, and select 50 industries of the testing data which is used to verify the aforementioned model. This study verifies: (a) the correct rate and return on investment from the original data (b) the correct rate of screening by decision tree and linear discriminant analysis (c) the return on investment by decision tree and linear discriminant analysis.