半導體產業對於各種高科技產業的景氣趨勢向來有領先指標的作用。 因此,本研究將以半導體股票作為分析股票和趨勢來創建投資組合的一個例子。本研究著眼於線性判別分析(Linear Discriminant Analysis, LDA)為數學統計的方法,將要分析的資料輸入至方法中,分析後可以得到一些線性的判別式,讓決策者容易進行決策。在大量資料未進行處理時,容易有雜訊資料(產生),此雜訊資料不僅容易混淆決策者進行決策,也可能造成誤判或漏判的跡象。因此,利用線性區別函數分析可以改善雜訊(資料問題)。 在本研究中,收集了不同研究機構的投資條件和歸因資料,對線性判別分析(LDA)進行了RESE分析,有助於找出長期投資並進行綜合評估。
The semiconductor industry has been the leader of indicators for a variety of high-tech industry boom trends. Therefore, this study will focus on semiconductor stocks as an example to analyze the stock and trends to create portfolio. It is easy to generate noise if the case of a large amount of data had not been processed. Not only will this problem confuse decision-makers in making their decisions, but also cause miscarriage of justice or missed data. The Rough Set Exploration System (RSES) can find the core factor to influence the decision. In this study, the investment conditions and attributed data of different research firms were collected to analyze the RESE with linear discriminant analysis (LDA), which helped find the long-term investment and make an overall assessment. The method of linear discriminant analysis (LDA) is then used as a classifier to generate the portfolio. Therefore, we use linear discriminant function analysis (LDA) to improve the problem of noise generated. The contribution of this study is that decision-maker can make the decision easily.