FORECAST THE DAILY INVESTMENT MARKET TREND USING A HYBRID MACHINE LEARNING TECHNIQUE

Abstract

Forecasting the movement of stock prices is vital for realizing the greatest possible return on an investment in stocks. It attempts to predict the path that stock prices will take in the future. Possible investors capitalize on the possibility of a crisis by redistributing their funds and turning prospective drawbacks into potential gains. But due to some losses in market trends, investors are hesitant to put their money into risky ventures. To overcome this issue, we proposed a Hybridized Random Support Vector Machine (H-RSVM) method. The purpose of H-RSVM method is to predict the daily market trend in investments. We collected the dataset of stock market investment and then collected data is preprocessed using min-max normalization. The preprocessed data is feature extracted using principal component analysis (PCA). As a result, our proposed H-RSVM method provides a superior performance in predicting the prices of stock market in terms of accuracy, precision, recall and f1-score measure.

Keywords- Stock market, Machine learning (ML), Investments, Prices, Funds, Hybridized Random Support Vector Machine (H-RSVM), Principle component analysis (PCA), Support vector machine (SVM), Random forest (RA).