IHSG CHANGE DETECTION DURING COVID-19 IN INDONESIA USING GAUSSIAN PROCESS

• Main Author: Arifani, Naqisya
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/86467
• Institution: Institut Teknologi Bandung

The COVID-19 pandemic has had a significant impact on Indonesia's stock market, particularly on the Indonesia Stock Exchange (IDX) or IHSG. This study aims to detect change points in IHSG stock prices during the pandemic using the Gaussian Process method. The research is motivated by the need to understand and analyze stock market fluctuations during crises, which can greatly influence investment decisions. The IHSG was selected for analysis as it reflects the overall condition of the Indonesian stock market. Weekly IHSG stock price data from 2019 to 2023 were analyzed using the Gaussian Process with Radial Basis Function (RBF) and Matern kernels. The analysis involved calculating the Generalized Likelihood Ratio Test (GLRT) to detect change points, with threshold variations of 10, 20, 30, 40, and 45 to assess detection sensitivity. The results identified thresholds of 35 for the RBF kernel and 20 for the Matern kernel as optimal for detecting change points, balancing sensitivity and specificity. Significant GLRT spikes, surpassing these optimal thresholds, were associated with major events during the COVID-19 pandemic in Indonesia. Conversely, low and stable GLRT values, below these thresholds, indicated normal market conditions without significant changes or high volatility. In conclusion, the Gaussian Process method with RBF and Matern kernels is effective in detecting significant changes in IHSG stock prices during the COVID-19 pandemic. This method aids investors in monitoring market volatility and planning adaptive investment strategies in response to market fluctuations during crisis like the pandemic.