Stock Forecasting using HMM and SVR

Stock markets are known for their highly volatile nature which makes it very complex. It changes continuously due to the effect of various attribute, of which many are unknown. This is why it is highly challenging to forecast future stock rates as no particular pattern is followed. However, in order to acquire profits, it is wise to keep in mind that factors should be analyzed and calculated before investing. Forecasting of stock prices is one of the challenging problems of dealing with models which are not of stationary form when trying to predict using machine learning. In recent times, many methods have been tried and tested in order to forecast the future rates of the stocks by keenly studying the previous rate trends. The Hidden Markov Model (HMM) is not a very familiar method that is used in predicting stocks. In this project, we'll be using HMM and Support Vector Regression (SVR) (comparatively more commonly used in stock forecasting) to predict stocks and compare their performance to deduce which model is more efficient. For this paper, public stock data of TESLA has been obtained from Yahoo Finance API.