Machine Learning is paving its way into every domain of our lives and especially in finance and economics. StockMarkets make for a lucrative avenue due to the opportunities it provides for investors to turn a profit if they can predict the trend of the market. In this Thesis, I used commonly available ensemble methods and compared it to stacked ensembles which involve creating a deep neural network for a meta learner. This meta learner will take input from our regressive models and try to eliminate any innate bias by learning the pattern between their predictions and actual stock prices. Upon testing with IBM stock, I have found out that using a well-tuned stacked ensemble has 2.59% better directional accuracy while compared to other ensemble methods while having a similar error percentage in prediction.