Strategi K-Medoids Clustering Berbasis Dynamic Time Warping Untuk Permasalahan Optimasi Portofolio

Agustina, Fikri (2025) Strategi K-Medoids Clustering Berbasis Dynamic Time Warping Untuk Permasalahan Optimasi Portofolio. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002211101-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Kesadaran investor terhadap keberlanjutan semakin meningkat, mendorong pertumbuhan investasi berbasis Environmental, Social, and Governance (ESG) sebagai pilihan utama di pasar modal. Dalam upaya mendukung perluasan praktik investasi berkelanjutan, Bursa Efek Indonesia meluncurkan indeks ESG IDX KEHATI yang berisi saham-saham perusahaan berkomitmen terhadap praktik bisnis berkelanjutan. Dalam hal ini, investor dituntut untuk tidak hanya mengejar imbal hasil yang tinggi, tetapi juga mengelola risiko dengan optimal melalui strategi diversifikasi yang efektif. Penelitian ini menerapkan pendekatan pengelompokan saham (clustering) untuk mengelompokkan saham berdasarkan kemiripan pola pergerakan harga sebagai dasar pembentukan portofolio optimal. Metode K-Medoids Clustering berbasis Dynamic Time Warping (DTW) digunakan karena kemampuannya mengatasi pergeseran waktu pada data deret waktu (time series data) yang umum terjadi dalam pergerakan harga saham. Saham dikelompokkan berdasarkan kemiripan pola harga, sehingga saham dalam satu klaster memiliki perilaku yang serupa. Diversifikasi dilakukan dengan memanfaatkan hasil klasterisasi agar portofolio mencakup saham dari berbagai kelompok, sehingga risiko tersebar dan dapat diminimalkan secara optimal. Hasil klasterisasi tersebut digunakan sebagai dasar optimasi portofolio dengan model Mean Variance Optimization (MVO). Hasil penelitian menunjukkan bahwa model MVO berbasis klasterisasi menghasilkan portofolio dengan risiko lebih rendah dan Sharpe ratio lebih tinggi dibandingkan model tanpa klasterisasi, sehingga memberikan keseimbangan yang lebih baik antara risiko dan imbal hasil. Pendekatan klasterisasi tidak hanya mendukung diversifikasi, tetapi juga mengurangi risiko konsentrasi pada saham yang berkorelasi tinggi, sehingga portofolio menjadi lebih seimbang dan terukur.
=========================================================================================================================================
Investor awareness of sustainability is increasing, driving the growth of investments based on Environmental, Social, and Governance (ESG) principles as a primary choice in the capital market. To support the expansion of sustainable investment practices, the Indonesia Stock Exchange launched the ESG IDX KEHATI index, which comprises stocks of companies committed to sustainable business practices. In this context, investors are expected not only to seek high returns but also to manage risks optimally through effective diversification strategies. This study applies a stock clustering approach to group stocks based on the similarity of their price movement patterns as the foundation for constructing an optimal portfolio. The K-Medoids Clustering method based on Dynamic Time Warping (DTW) is employed due to its ability to address time shifts commonly found in time series data of stock price movements. Stocks are grouped according to the similarity of their price patterns, ensuring that stocks within the same cluster exhibit similar behavior. Diversification is carried out by leveraging the clustering results so that the portfolio includes stocks from various groups, thereby spreading and minimizing risk optimally. The clustering results are then used as the basis for portfolio optimization using the Mean Variance Optimization (MVO) model. The findings show that the clustering-based MVO model produces portfolios with lower risk and higher Sharpe ratios compared to models without clustering, providing a better balance between risk and return. This clustering approach not only supports diversification but also reduces concentration risk in highly correlated stocks, resulting in a more balanced and
measurable portfolio.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	K-Medoids, Dynamic Time Warping, Mean-Variance, Optimasi Portofolio, ESG IDX KEHATI, K-Medoids, Dynamic Time Warping, Mean-Variance, Portfolio Optimization, ESG IDX KEHATI
Subjects:	H Social Sciences > HA Statistics > HA30.3 Time-series analysis H Social Sciences > HG Finance H Social Sciences > HG Finance > HG4529.5 Portfolio management H Social Sciences > HG Finance > HG4910 Investments Q Science > QA Mathematics Q Science > QA Mathematics > QA278.55 Cluster analysis
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Fikri Agustina
Date Deposited:	31 Jul 2025 02:26
Last Modified:	31 Jul 2025 02:26
URI:	http://repository.its.ac.id/id/eprint/123785

Actions (login required)

View Item