Stock Portfolio Optimization of IDX30 using Agglomerative Hierarchical Clustering and Ant Colony Optimization Algorithm

Abstract

The stock market offers high profit opportunities but also entails significant risks, making portfolio optimization essential to help investors manage risk and maximize returns. This study aims to cluster IDX30 stocks to form a more diversified portfolio, determine the optimal stock weights, and evaluate portfolio performance. The method employed is Agglomerative Hierarchical Clustering (AHC) with Ward linkage for clustering stocks based on financial ratios, with the silhouette score used to evaluate cluster quality. Subsequently, the Ant Colony Optimization (ACO) algorithm is applied to optimize stock weights in the portfolio based on the clustering results. The findings indicate that the best portfolio is obtained in clusters 5 and 6, with a maximum fitness value of 0.064555 and a portfolio return of 0.000814. Portfolio performance evaluation using the Sharpe ratio yields a value of 0.044767 for both clusters, indicating that the resulting portfolios are efficient. This research is expected to contribute to the development of more accurate and practical data-driven investment strategies for investors.