I have an idea to use so called "Depth of Market" in stock trading to predict price. I downloaded data from "Depth of Market" for several months and pre-calculated it , so now in dataset you can see: Opentime - current time parameters, Price_1 and Price_2 - current instrument price (Bid and Ask), AVG_Volume_1 - is average Volume "Above" current price, AVG_Volume_2- is average Volume "Below" current price, STDEV_Volume_1 - is standart diviation "Above" current price, STDEV_Volume_2 - is standart diviation "Below" current price F_Price_1 and F_Price_2 - future instrument price (Bid and Ask) (just nearest 20step movement from current), price_more - is flag 1 or 2 : is future movement is "Up" or "Down" I am new in Machine Learning World so it is black box for me , so I simple try to use some models to predict price, but result is it good yet, and it will be cool to see what maybe other ways \models can do. It is intersting to see different variants, the task maybe impossible because of little number of features , sorry . But still awesome.