I am doing a university project on index/stock price prediction. I plan to use a combined cnn-lstm model, and I have several different types of data: Open High Low Close Volume, values, fundamental data such as unemployment and various rates, technical indicators like RSI, MACD and others, and moving averages like SMA, EMA, WMA and etc. At this moment I am using the following transformations

  • for these OHLCs - simple differentiation
  • for fundamental data - logarithmization
  • for moving averages - subtract the candle opening value from the value of this moving average
  • indicator values unchanged

Then I use StandardizeNormalizer for all dataset. I also tried normalizing (robust scaling, standardization, minmax scaling too) each sequence separately, and differentiating all the data, but it was not effective And i think my current approach is not effective either. What is the best way to prepare this type of data for the network?



You must log in to answer this question.