Strategy Description:
DeepNet is basically a Delta-One predictive trading strategies, which continuously estimate the probability spectrum of future price movements of an underlying instrument, and to make trade decisions that yield the highest potential profit. It is derived from a customized Bayesian Network Topology with unsupervised machine learning techniques to keep trace on market regime change. The model keep continuous learning and are dynamically updating to adapt to the most recent market situation.
This trading system, currently applied to Foreign Exchange Market, has the following trading behaviors, features and requirements:
- It is a kind of mid-to-high frequency trading strategy.
- Holding period can be ranging from minutes to days.
- There could be as many as tens trades per day.
- It might carry positions overnight.
- Capital size shall be at least as US$10k or equivalent.
- Implement on trading account with at least 50:1 leverage ratio.
