I have 365 dataframes each containing 30 rows. Each row represents a day in the future.Dataframes contain 4 features and target. The main feature is the amount of orders for each day for the next 30 days.Target is the actual sales. Other features are not relevant here.
What would be the best model for this purpose, where the aim is to predict the future sales,when we know how much we have orders for next 30 days and we have other 4 features available as well.
Can this be done somehow with Randomforest/Gradboost etc. so that the model would learn the behaviour?I guess the data has to be in a single dataframe for these types of models?
Or should this be done with some type of neural network? And what type of neural network would suit best?
Any python examples of how the input data should be organized and what type of model should work are welcome!
EDIT!
As an example, for dates 2016-03-03 and 2016-03-04 the head of the dataframes looks like this:
