Study on Machine Learning Techniques In Financial Markets

Portfolios, securities, stock market forecasting, risk management, debt management are all important pillars of the financial world. These pillars rely heavily on adequate and accurate prediction. These are problems that, on the small scale, affect individuals and their financial conditions and, on the larger scale, can be detrimental to a country's financial well-being. This paper, is an attempt at understanding the various algorithms and platforms involving Machine Learning in Financial Markets and reaching informed conclusions on parameters like accuracy, efficiency, speed and usability. In this paper, primarily, different trading techniques are introduced and their effectiveness in quantitative trading and, in general, finance to generate alphas is observed. These techniques, as observed, are categorised by their reliance on Neural Networks, Support Vector Machines and other quantitative variables in finance. Classifications on the basis of supervised and unsupervised techniques and K-Mean clustering are also made. Further, this paper also delves into the hitherto unpredictable and unmovable phenomena in market and public psychology and attempts to suggest a viable solution to it.