Optimasi Portofolio Saham Di Bursa Efek Indonesia Dengan Simulasi Monte Carlo Dan K-Means Berbasis Jarak Dynamic Time Warping (DTW)

Nurrifki, M. Iqbal (2026) Optimasi Portofolio Saham Di Bursa Efek Indonesia Dengan Simulasi Monte Carlo Dan K-Means Berbasis Jarak Dynamic Time Warping (DTW). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221061-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (10MB) | Request a copy

Abstract

Pertumbuhan pesat jumlah investor di Bursa Efek Indonesia meningkatkan urgensi penerapan strategi investasi berbasis analisis kuantitatif dan pengelolaan risiko yang terukur. Penelitian ini mengintegrasikan analisis fundamental, klasterisasi K-Means berbasis jarak Dynamic Time Warping (DTW), dan simulasi Monte Carlo untuk membentuk portofolio saham yang optimal. Seleksi awal dilakukan terhadap 890 emiten menggunakan lima kriteria fundamental secara simultan, yaitu Price to Book Value (PBV) kurang dari 1, Return on Equity (ROE) lebih dari 8%, Net Profit Margin (NPM) lebih dari 6%, Return on Assets (ROA) lebih dari 4%, dan Moving Average (MA) 50 hari bernilai positif, sehingga diperoleh 34 saham terpilih. Saham-saham tersebut dikelompokkan menggunakan algoritma K-Means berbasis jarak DTW, menghasilkan 4 klaster optimum dengan nilai Silhouette Score sebesar 0,6347 yang mengindikasikan kualitas pengelompokan berada pada kategori cukup baik, dan dikonfirmasi secara statistik melalui PERMANOVA. Berdasarkan visualisasi pola pergerakan harga, Klaster 2 yang beranggotakan saham ADMF, ITMG, TKIM, dan UNIC dipilih karena menunjukkan pola pergerakan harga yang relatif stabil dibandingkan klaster lainnya. Optimasi portofolio dilakukan menggunakan simulasi Monte Carlo dengan 10.000 iterasi pada dua skenario, yaitu skenario per tahun (2024 dan 2025) serta skenario gabungan dua tahun (2024-2025). Skenario 2025 menghasilkan kinerja terbaik dengan Sharpe Ratio sebesar 0,8160, diikuti skenario 2024 sebesar 0,4192, dan skenario gabungan sebesar 0,3490. Evaluasi durasi efektivitas bobot pada periode out-of-sample 1 Januari-6 Februari 2026 menunjukkan bahwa bobot skenario 2025 dan gabungan efektif selama tiga minggu pertama, sedangkan bobot skenario 2024 tidak efektif sejak minggu pertama. Temuan ini menunjukkan bahwa integrasi analisis fundamental, klasterisasi K-Means DTW, dan simulasi Monte Carlo mampu menghasilkan portofolio yang efisien dalam menyeimbangkan risiko dan return, dengan bobot optimal yang perlu dikalibrasi ulang secara berkala menggunakan data terbaru agar kinerjanya tetap relevan dengan kondisi pasar yang berlaku.
====================================================================================================================================
The rapid growth of investors and trading activity in the Indonesia Stock Exchange has increased the urgency of implementing investment strategies based on quantitative analysis and measurable risk management. This study integrates fundamental analysis, K-Means clustering based on Dynamic Time Warping (DTW) distance, and Monte Carlo simulation to form an optimal stock portfolio. An initial screening was conducted on 890 listed companies using five fundamental criteria simultaneously, namely Price to Book Value (PBV) less than 1, Return on Equity (ROE) greater than 8%, Net Profit Margin (NPM) greater than 6%, Return on Assets (ROA) greater than 4%, and a 50-day Moving Average (MA) with a positive value, resulting in 34 selected stocks. These stocks were then clustered using the K-Means algorithm based on DTW distance, yielding an optimum of 4 clusters with a Silhouette Score of 0.6347, indicating a good clustering quality, and statistically confirmed through PERMANOVA. Based on the visualization of price movement patterns, Cluster 2 consisting of ADMF, ITMG, TKIM, and UNIC stocks was selected due to its relatively stable compared to the other cluster. Portfolio optimization was performed using Monte Carlo simulation with 10,000 iterations under three scenarios, namely the annual scenarios (2024 and 2025) and the combined two-year scenario (2024-2025). The 2025 scenario yielded the best performance with a Sharpe Ratio of 0.8160, followed by the 2024 scenario at 0.4192, and the combined scenario at 0.3490. The evaluation of optimal weight effectiveness duration during the out-of-sample period of January 1–March 31, 2026 showed that the weights from the 2025 and combined scenarios remained effective for the first three weeks, while the weights from the 2024 scenario were not effective from the first week. These findings indicate that the integration of fundamental analysis, K-Means DTW clustering, and Monte Carlo simulation is capable of producing a portfolio that is efficient in balancing risk and return, with optimal weights that need to be periodically recalibrated using the latest data to ensure portfolio performance remains relevant to prevailing market conditions.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Dynamic Time Warping, K-Means, Monte Carlo, Optimasi Portofolio, Durasi Efektivitas Bobot, Dynamic Time Warping, K-Means, Monte Carlo, Portfolio Optimization, Stocks, Weight Effectiveness Duration
Subjects:	H Social Sciences > HG Finance > HG4529.5 Portfolio management
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	M. Iqbal Nurrifki
Date Deposited:	28 Jul 2026 04:04
Last Modified:	28 Jul 2026 04:04
URI:	http://repository.its.ac.id/id/eprint/138335

Actions (login required)

View Item