CryptoNet: Using Auto-Regressive Multi-Layer Artificial Neural Networks to Predict Financial Time Series

When analyzing a financial asset, it is essential to study the trend of its time series. It also necessary examine evolution and activity over statistically analyze possible future behavior. Both retail institutional investors base their trading strategies on these analyses. One most used techniques series dynamic structure using auto-regressive models, simple moving average models (SMA), mixed (ARMA). These techniques, unfortunately, do not always provide appreciable results both at statistical level as Risk-Reward Ratio (RRR); above all, each system has pros cons. In this paper, we present CryptoNet; based extraction exploiting vast potential artificial intelligence (AI) machine learning (ML). Specifically, focused trends by developing an neural network, trained tested two famous crypto-currencies: Bitcoinand Ether. CryptoNet algorithm improved classic linear regression model up 31% MAE (mean absolute error). Results from work should encourage in sectors classically reluctant adopt non-standard approaches.