README.md

Hello, This Notebook Contains Machine-Learning Service To Training Multioutput Regression Model.

- the produced model is a piece of complement project named(Reactions Prediction).

- We use multioutput regression technique to predict all values by one model insted of train one model per category (there is also mathmatical reasons beyond using it).

- We use Scikit-Learn library for training and testing models.

- We implement different techniques for feature selection, and we use Grid_Search technique for parameters tunning.

- Finally we export the best model into .pickle file