Advertisement
The GenDeep Learning AI (By muratyazici)
The GenDeep Learning AI Discussion
The GenDeep Learning AI!!
My name is Murat Yazici. I am a PhD candidate in Statistics.I have several papers, two books, book chapters, and conference presentations about modeling worldwide.I invite you to follow the system of my EA called the GenDeep Learning AI.The system targets 20% monthly growth with respect to max. 8% DD for each trade.The EA opens only one trade simultaneously to manage the Risk.
The GenDeep Learning AI is based on a novel trend prediction algorithm developed by me.The algorithm includes some AI approaches like Reinforcement Learning and Fuzzy Statistics.
We know that the market is not deterministic. It is stochastic. So, to get the perfect trading model seems impossible.However, we can obtain a High Win-Rate and decrease the Risk as much as possible.The EA has solved the following optimization problem related to its novel trend prediction algorithm with any fixed lot size.@What is the optimum profit for each trade while the Win-Rate is increasing and the Risk is decreasing?
The GenDeep Learning AI is a scalping strategy. It has hard coded take profit and stop loss points.If there is a loss, the EA reduces the risk till recovering the loss. To recover the loss, the EA needs only 10 profitable trades.
__________________Some current statistics:_____________________Win rate : 83.67% (41 profitable out of 49 trades)Profit and Recovery Factors: 1.63 and 1.64Avg. holding time : 3 hoursAvg. Profit : 751.96 USDExpected Payoff : 243.88 USD/trade_________________________________________________________
In general, the EA opens avg. 5 trades each day. It runs on EURUSD, GBPUSD, USDCHF, and AUDUSD pairs.The system completed its two weeks with 11.95% growth, as in the table.I will share its weekly performance here.See you next Friday!
Murat Y.
Hi muratyazici,
The use of machine learning and AI for trading is fascinating and your system sounds very interesting. I remember listening to the "The Quantopian Podcast" and they discussed possible uses for machine learning in quant trading. In my personal experience I've found it very difficult to prevent the ML models from overfitting the data and learning from the noise of the financial markets. Currently I've been experimenting with algo trading based on rather simple rules but I want to add ML to optimise the parameters of the algorithm based on backtest training.
Thanks for sharing your strategy and I look forward to following your results.
