Identifying Homogeneity of Small-Cap Stocks in Indian Market: A Data Mining Approach Shuvashish Roy*, Rajib Bhattacharya** * Financial Advisor, Hazrat Khajar Bashir Unani Ayurvedic Medical College & Hospital Foundation, Jamalpur, Bangladesh. Email: shuvashishroy@gmail.com ** Assistant Professor, Institute of Business Management, The National Council of Education Bengal, Kolkata, West Bengal, India. Email: rajib.conference@gmail.com Abstract Investors in equity shares look for two basic aspects while investing i.e. consistently rising returns with a decreasing or at least stabilized level of risk involved. Amidst the numerous stocks available in the market which differ widely on the basis of different aspects i.e. segment, sector, industry, market capitalization etc. it becomes a challenge for the investor to form a diversified portfolio where heterogeneity of the constituent stocks is the main criterion. Thus it is imperative that the basis be finalized on which the heterogeneity of the stocks shall be determined. Traditionally portfolios have been constituted on the basis of low coefficient of correlation of returns from the constituent stocks. The dissimilarity of co-movement of returns from stocks has traditionally been attempted to be maximized in portfolios. Another method of grouping similar stocks by using data mining approach is fast gaining popularity. This approach uses clustering technique to group homogeneous stocks on the basis of a set of selected criteria. These criteria can be financial ratios, indices or any other related matrices. Advanced versions of this technique can group homogeneous time series data as well. This paper attempts to identify homogeneous clusters of companies in the Indian small-cap segment based on valuation ratios. Valuation ratios have been selected to be the grouping criteria as these were not been used in earlier studies by researchers and scholars. The small cap companies in India have been chosen for this study for its better resilience and recovering potential compared to mid cap and large cap companies. The companies constituting the CNX NIFTY Small Cap 50 Index have been considered in the study. International Journal of Business Analytics and Intelligence Submitted: 05 January, 2019 7 (1) 2019, 53-63 Accepted: 20 February, 2019 http://publishingindia.com/ijbai/ Introducton A system of categorization of stock market would be useful to investors and fnancial analysts, providing them with the opportunity to predict the stock price changes of a company vis-a-vis other companies. In recent years, clustering companies in the stock markets based on their similarities on different aspects has increasingly become a common practice. However stock price data are high-dimensional in nature and the changes in the stock price usually occur with shift, which makes the categorization a complex issue. Clustering Method is an adaptive procedure in which homogeneous objects are clustered or grouped together, based on the principle of maximizing the intra-class similarity and minimizing the inter-class similarity. It is essentially a data mining technique in which similar data are automatically placed into related groups without advanced knowledge of the group defnitions. Clustering of companies in the stock market is very useful for managers, investors and policy makers. There are numerous companies listed in the Indian market which vary widely based on a host of aspects i.e. the industry, size, capitalization, business models etc. However, the investors have only two aspects to consider i.e. consistently increase returns with a cap on risk involved. Earlier studies have used fnancial ratios as clustering factors. To capture the market price of the stocks, which is of prime importance to the investors, Keywords: Cluster Analysis, Valuation Ratios, Small Cap Sector, CNX NIFTY Mid Cap 50 Index