ABSTRACT
Cryptocurrency has become a global phenomenon in the financial sector, and it is now one of the most widely traded financial assets on the planet. Cryptocurrency is regarded as a confusing topic in finance due to its significant volatility. It is not just one of the most sophisticated and arcane fields among financial products. The key issue in this development is the fast pace of Cryptocurrency swings. The authors have proposed a number of machine learning strategies for resolving the above mentioned issue. The proposed technique would be a great choice to anticipate and forecast Cryptocurrency prices using the index and constituents. The comparison of the proposed mechanism with the existing methods is also accomplished. In this paper, two models have been proposed named as Light Gradient-boosting Machine (LGBM) and Auto-regressive Integrated Moving Average (ARIMA). It is witnessed that the ARIMA model gives more accurate results while forecasting values, whereas LGBM used lag values and hyper parameters based on closing variables with 7-fold cross validation technique to train the model.
