Development clustering system IDX company with k-means algorithm and DBSCAN based on fundamental indicator and ESG

Area of Science:

• Finance
• Data Science
• Machine Learning

Background:

• The COVID-19 pandemic created investment opportunities but exposed weaknesses in traditional stock market indices.
• Existing indices and classifications like 'blue chips' may not accurately reflect companies' resilience or future potential.
• Index formulation and constituent selection can be biased, leading to suboptimal investment guidance.

Purpose of the Study:

• To address biases in stock index composition and company classification.
• To leverage machine learning for objective clustering of companies based on fundamental and ESG data.
• To identify distinct investment clusters within the Indonesian Stock Exchange (IDX) using the KOMPAS100 index.

Main Methods:

• Utilized K-Means and DBSCAN machine learning algorithms for clustering.
• Employed fundamental indicators and Environmental, Social, and Governance (ESG) attributes from the KOMPAS100 index dataset.
• Applied data-driven techniques to group companies objectively.

Main Results:

• Successfully clustered companies within the KOMPAS100 index into five distinct groups.
• The clustering was based on a combination of fundamental financial indicators and ESG performance.
• The results suggest a more nuanced view of company performance beyond traditional index classifications.

Conclusions:

• Machine learning offers a robust method to counter human bias in company classification and index composition.
• The identified clusters provide valuable insights for novice investors navigating market volatility.
• Objective, data-driven clustering based on fundamental and ESG factors can enhance investment strategies.