Portfolio Generation for Indian Stock Markets Using Unsupervised Machine Learning

Abstract

Portfolio Management is a concept of selecting the proportions of various assets that is to be held in a portfolio to have a good return without a significant risk exposure. Portfolio optimization is one important building block in financial management and investment banking. One possible strategy for minimisation of risk is by enlarging or varying its field of operation for the portfolio. Constructing an optimal portfolio by judging and selecting the best possible combinations of different portfolio is a computationally challenging problem since it comes up with an exponential complexity. Here, we have proposed a simple k-means based clustering strategy for an optimal portfolio. BSE100 stocks are represented by their fundamental financial ratios. Clustering is performed, and a prototype stock is selected from each of the clusters. An equal investment strategy demonstrates superior return as compared to the indices as well as top mutual funds. The classification is done by considering a host of investment parameters. We compare the rate of return of these stocks to the benchmark of Indian Stock Exchange.