| Summary: | 碩士 === 國立中興大學 === 財務金融學系所 === 103 === This paper adopts the Apriori Algorithm of Big Data methods to calculate and in-vest in Taiwan stock market. First of all, we construct four types of sample by “Today’s daily returns are 6% and next day’s returns are 6%”, “Today’s daily returns are 6% and next day’s returns are -6%”, “Today’s daily returns are -6% and next day’s returns are -6%” and “Today’s daily returns are -6% and next day’s returns are 6%” for calculating to find rules. In empirical results, the “Today’s daily returns are 6% and next day’s re-turns are 6%” can only find out rules. After rules calculated by Apriori Algorithm, our investment performance can earn the positive accumulative annual return by rules in Taiwan stock market. Furthermore, compare with Taiwan’s market index, previous three years have positive abnormal return. Finally, compare to the strategy of benchmark by Mean-Variance model. The annual returns of Apriori Algorithm’s rules can beat the method of benchmark. We surmise the Mean-Variance model have the problem of standard error, hence the performance of Mean-Variance model is worse than Apriori Algorithm’s rules.
|
|---|
