Identifying optimistic stocks with K-means clustering algorithm

Selecting stocks from a large number of active stocks is a critical and challenging investment decision due to high volatility and biased decision-making. The abundance and availability of financial data provide machine learning an advantage to optimise investment decision processes. The $k$-means algorithm of machine learning is used to cluster observations into different groups where each group contains observations with similar properties. In this paper, a risk-managed momentum strategy is proposed to identify promising stocks. We employ five risk-adjusted factors to cluster stocks and select the clusters with the best-performing stocks for equity portfolio construction. It enhances the fundamental investment decision of stock selection to construct optimised portfolios. The proposed strategy remarkably outperforms the standard momentum technique and the stock market indices. It reduces transaction costs and hedges investors during severe financial crises. An outstanding feature of the proposed method is the superior risk-adjusted performance with a smaller number of stocks, which is indispensable for individual investors who seek fewer stocks for investment.