A Novel LSTM based Approach for Crypto Currency Price Prediction

Abstract

The nature of cryptocurrencies is decentralized and they have potential of large returns, due to this nature of crypto currencies, they have increased in approval in form of investment. Due to the erratic and volatile nature of the crypto currency market, it can be difficult to predict their pricing. As a result, reliable price forecasts are essential for investors to make wise investment choices. The proposed LSTM based approach will create machine learning models using open-source libraries like pandas, NumPy, and Scikit-learn. Cross-validation will be utilized to test the working of different models, and the one gives best output will be taken as the final model. The aim of this research is to create a machine learning algorithm that can forecast bitcoin values. Predicting the future price swings of crypto currencies like Bitcoin and Ethereum has become a crucial research subject as their use and popularity have increased. By using regression and deep learning algorithms, the work seeks to use different number of machine learning models to analyze historical bitcoin data and forecast future prices. The end result of this work will be a very effective and accurate model with R2 score of 0.96 for train data and 0.97 test data for forecasting crypto currency values, which traders, investors, and researchers may use to decide wisely on investing in crypto currencies.