Application of Data Mining Techniques for Stock Investment – A case study of Indonesian Exchange Traded Fund (ETF R-LQ45X)
| Autor: | RIBKA ULI, 雷蓓可 |
|---|---|
| Rok vydání: | 2019 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 107 This thesis is an application of a research framework of data mining on stock investment proposed by a recent study (Wu, Chung, and Hung 2019). We apply this research framework to the case of R-LQ45x. The daily data is selected for 11 years, from 2008 to 2018. The purpose of this study is to help the mutual funds managers improving their investment decision about buying and selling or holding stocks to get more profit. The problems that involves deciding the right time to buy, sell and hold stocks are feature selection, labeling and time series (long term or short term). Labeling was used to define buy signal and sell signal. Buying and selling signal was made by developing two binary classifiers (buy or hold and sell or hold). Long term means buying and selling every time buying and selling signal is emitted and short term is similar with long term but if it doesn’t meet sell signal, it must be sold out on the last day of the quarter. The method of this study starts from calculating technical analysis as benchmark, then labeling, finding the most accurate algorithm as data mining predictor and comparing Return On Investment (ROI) with the best data mining predictor and the benchmark. The final result show that these data mining techniques win 53 quarters (out of 72) in short term performance and win 5 quarters (out of 6) with 21.47% in long term performance. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|