I need a Machine Learning Engine to Optimize entry and and exit . I have programmed a custom indicator for the Ninatrader platform that will output different up to 16 variables (Characters) from a set of 6 possible outcomes. To simplify I have made them Characters, ABCXYZ.. In a series (look back period) of up to 10.. Example of one series of a look back of 10. There will be 8 of these.
In essence I would like to it to recognize when which pattern of letters emerges, whether it would be a good time to buy or exit the market. I could be optimized for when a certain pattern emerges there is a n% probability it will go up or go down.
I have included an image of the indicator in use within Ninja trader so to better understand. Each color corresponds to a letter. I also have attached a longer explanation of my requirements and some sample data.
I would like it to work with Ninja trader but am flexible and it can be a separate application, if I can import to it a csv with the values of the variables, bar number and bar last price. . I can provide all the data necessary.
Open to ideas and suggestions.