e-ISSN: 2582-5208 International Research Journal of Modernization in Engineering Technology and Science ( Peer-Reviewed, Open Access, Fully Refereed International Journal ) Volume:03/Issue:10/October-2021 Impact Factor- 6.752 www.irjmets.com www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science [1316] A COMPARATIVE STUDY OF SVM, LSTM AND LR ALGORITHMS FOR STOCK MARKET PREDICTION USING OHLC DATA Isha S Meshram *1 , Prajakta J Kulal *2 *1 Dept Of Chemical Engg., Vishwakarma Institute Of Technology, Pune, Maharashtra, India. ABSTRACT Prediction in the stock market is challenging and tricky process for financial investors. Many researchers have performed various studies to sense the future market developments. The variation and complex inter- dependency of multiple parameters make stock market predictions as a difficult outcome. Machine Learning have been proven to be helpful in such cases to forecast the stock market values. The parameters involved and the commonly used models are discussed and compared in this paper. A comparative study of algorithms: Scaled Unscaled LR, SVM, LSTM are discussed. In machine learning, LR is a fundamental approach by which a linear trend can be obtained. But SVM have modern features such as high exactness and certainty. LSTM is a kind of recurrent neural network (RNN)and uses an appropriate gradient algorithm. In this paper algorithms were compared to predict stock values. The methodologies and results of the algorithms were performed using python codes which are discussed further in this paper. Keywords: Stock Market Predictions, AI Methods, ANN, LSTM, SVM, LR, Python. I. INTRODUCTION The Stock marketplace refers to the gathering of markets and exchanges in which ordinary sports of buying, selling, and issuance of stocks of publicly-held corporations take place. It is essential to perform a complete study of Stock Market earlier than making a funding. While buying a share, definitely few quantities are bought of the enterprise with an expectation to make an income if there's a boom withinside the enterprise value. Even before shopping for a fabric or phone, investigation of their first-rate and overall performance are also being analyzed. Similarly, while taking a decision of buying stock , then definitely predictions are necessary so that earned hard cash is invested withinside the proper area. If successful, this can be useful for investors to make a positive contribution to the economy. As every algorithm has its own advantage and disadvantage this research will study different machine learning techniques commonly use for stock market prediction. This paper demonstrates a comparative study to find out optimal technique for stock market prediction under different circumstances. The main purpose of predictive model is to minimize error and to obtain maximum accuracy. There is no such optimal solution for stock market prediction as there is no specific technique to analyze the exact values. ANN is employed as a random function approximation tool. These sorts of tools help estimate the foremost cost-effective and ideal methods for arriving at solutions while defining computing functions or distributions. ANN takes data samples instead of entire data sets to reach solutions, which saves both time and money. ANNs are considered fairly simple mathematical models to reinforce existing data analysis technologies. For predicting stock prices, the Artificial Neural Network (ANN), an area of Artificial Intelligence (AI), is a preferred way to identify dark and hidden patterns in data that are suitable for stock market prediction, and it also has its wide applications in various domains such as stock price prediction, developing security trading systems, predicting bond ratings, modeling foreign exchange markets and forecasting financial distress, evaluating default risk for applications of loans, mortgages and credit cards and much more. Further ANN find their applications in various fields like in image processing for face recognition , face detection , Image recognition and image reconstruction, Noise removal. In health care sector for disease recognition , Computer Aided surgery. In defense for automated target recognition, autonomous soldier robot. ANN applied in automobile field too for self-driving cars. THEORY In order to forecast the stock prices of SBIN.NS and to detect profits, reduce risks three algorithms are focused in this paper which are discussed below.