Clustering Stock Prices of Financial Sector Using K-Means Clustering With Dynamic Time Warping

Abstract

An investor is a person who invests money in a business to make for a profit. Investment instruments in the capital market include stocks, bonds, warrants, rights, mutual funds, and various other derivative instruments. According to the IDX, the number of stock investors has increased to 3,988,341 SID as of June 24, 2022, an increase of more than 536 thousand SID or 15.6% from the previous year. Every investor wants to benefit from the shares they own. So it is necessary to consider which groups have desired price fluctuations. In this study, data on the share prices of financial sector companies are used for the period April 1, 2021, to March 31, 2022. The variables used are open, close, and HML (High Minus Low) stock prices. The method used is K-Means clustering with Dynamic Time Warping (DTW) distance. The K-Means was chosen because it is commonly used for large data scales; besides that, K-means with DTW was chosen because it is a non-linear sequence alignment distance, so it is considered suitable to be applied to stock price data in the form of time series data. The analysis was carried out by comparing the results of K-means with Euclidean and DTW distances. It was concluded that the DTW distance gave a higher silhouette score than the Euclidian distance.