Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/166231
Title: Pembentukan Portofolio menggunakan K-Means Clustering Berbasis Minimum Spanning Tree dan Beberapa Variasi Pembobotan
Other Titles: Portfolio Building using K-Means Clustering Based on Minimum Spanning Tree and Several Weighting Variations
Authors: Septyanto, Fendy
Budiarti, Retno
FRIANDY, MUSTIKA DEWI
Issue Date: 2025
Publisher: IPB University
Abstract: Dalam investasi saham, tantangan utama adalah bagaimana membentuk portofolio yang mampu memberikan keuntungan optimal dengan risiko yang terkendali. Penelitian ini bertujuan untuk membentuk portofolio saham yang optimal dengan menggabungkan metode k-means clustering berbasis Minimum Spanning Tree (MST) dan tiga teknik pembobotan, yaitu metode Markowitz, Inverse Degree Centrality Portfolio (IDCP), dan maximum Sharpe ratio. Proses dimulai dengan menghitung korelasi antar saham dalam indeks LQ45 untuk membentuk graf korelasi antar saham. Graf ini kemudian disederhanakan menjadi struktur MST untuk menangkap hubungan utama antar saham. Berdasarkan struktur ini, saham dikelompokkan menggunakan k-means clustering. Dari masing-masing kelompok, dipilih satu saham terbaik berdasarkan nilai Sharpe ratio. Saham-saham terpilih kemudian dibobot menggunakan tiga pendekatan berbeda. Hasil penelitian menunjukkan bahwa metode pembobotan maximum Sharpe ratio menghasilkan portofolio dengan kinerja terbaik, baik dari sisi return maupun efisiensi risiko. Temuan ini menunjukkan bahwa kombinasi analisis jaringan, teknik klasterisasi, dan pembobotan berbasis efisiensi return-risiko dapat membantu investor membentuk portofolio yang optimal dan terdiversifikasi dengan baik.
In stock investment, the main challenge is how to form a portfolio that is able to provide optimal profit with controlled risk. This study aims to form an optimal stock portfolio by combining the Minimum Spanning Tree (MST)-based K-Means Clustering method and three weighting techniques, namely the Markowitz method, Inverse Degree Centrality Portfolio (IDCP), and maximum Sharpe ratio. The process begins by calculating the correlation between stocks in the LQ45 index to form a connected graph between stocks. This graph is then simplified into an MST structure to capture the main relationships between stocks. Based on this structure, stocks are grouped using k-means clustering. From each group, one best stock is selected based on the Sharpe ratio value. The selected stocks are then weighted using three different approaches. The results show that the maximum Sharpe ratio weighting method produces the best performing portfolio, both in terms of return and risk efficiency. The findings suggest that a combination of network analysis, clustering techniques, and risk-return efficiency-based weighting can help investors form an optimal and well-diversified portfolio.
URI: http://repository.ipb.ac.id/handle/123456789/166231
Appears in Collections:UT - Actuaria

Files in This Item:
File Description SizeFormat 
cover_G5402211007_853ab08976724414a2c38a824755aa22.pdfCover159.91 kBAdobe PDFView/Open
fulltext_G5402211007_9da6db84c9b54d3eadb98bbb2ac84c63.pdf
  Restricted Access
Fulltext481.55 kBAdobe PDFView/Open
lampiran_G5402211007_5c2c96221d5f4e55add6e78b5b18bcfa.pdf
  Restricted Access
Lampiran2.28 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.