Select data feeds (your neural network's input data)
Classic time-series forecasting based on historical price data.
We collect and analyze multiple news sources in real time, and compare the effect of the general sentiment across channels with corresponding price swings.
A data spurce generated from 1000 crypto related keywords and compared to historical price swings. Normalized sentiment output is generated by a separate internal ML algorithm.
This data source is created by following the 10,000 largest on-chain address, the flow between them and corresponding historical price swings.
This data source is created using bots which follow twitter, Wechat, Telegram and other public channels, to come up with an average "crowed sentiment" input.
Using data from OTC partners, professional traders and analysts, the data source provides an average of the sentiment of industry professionals.
Select trading strategy