Investigation of Big Data Analytics for Forecasting Cryptocurrency Value Patterns

The year 2017 was an interesting year for Bitcoin investors in general as well as the activist of blockchain. The presence of similar coins utilizing smart-contract technologies such as Ethereum increases the popularity of Bitcoin as one of the first application of blockchain technology. The highly volatile price movement of cryptocurrency makes almost every prediction using technical analysis and fundamental analysis from experts do not have high accuracy in practice. Experts are still very difficult to predict cryptocurrency price movements in general and bitcoin specifically for short and medium term. Most predictions made by experts still use fundamental analysis and market predictions for the long term. Charlie Lee as the creator of Litecoin predicts that in the long run, the future of the cryptocurrency is very bright but for the short term nobody knows. This study aims to provide an overview to experts and the public about the future predictions of cryptocurrency prices in general through the methods of Big Data analysis approaches in the form of neural networks. Various considerations such as determining features, variables, classes and forms of data processing used are discussed in later chapters of this research. The overall result shows good result from 7 out of top 10 cryptocurrencies by market capitalization being bullish until the end of Q4 2019. By using 19 different datasets, we obtained 8 results with high confidence in forecasting the price in the future.