Classification of Stocks according to Future Profitability using Unsupervised Learning

Abstract

Selection of stocks that are suitable for an investment is a complex task. The main aim of every investor is to earn the maximum possible returns on investment. As a result stock market analysis mechanisms such as technical analysis has gained prominence in short term trading. Neural network models for stock market forecasting have been successfully used and are gaining importance in the financial sectors. Neural network models have proven to be far more accurate in providing accurate results for the stock market domain compared to other models. This research proposes a novel approach in using an unsupervised neural network model, known as Self Organizing Maps for the purpose of classification of stocks according to future profitability. The research initially extracted technical indicators as features and uses a Self Organizing Map to classify stock according to how they perform in the future. The empirical study carried out provided a significant reduction of the quantization error providing the ability to identify clusters based on a profit labeling criterion.