Regression module may be useful for u.

I ask you how many varibales in your module (depenedent & independent)

you can use Multiple regression or multivarite regession and and model give a high R2 is useful >

Unfortunately for cases like this, we still witness "artezanal" Data Mining and Data Science proces, that is, any direct opinion, as still usual in our area, will seem alchemy. For a more targeted alchemy, I suggest a controlled experiment with the 5 best published algorithms.

The Linear Multiple Regression is the first step. It is the grandfather of the artificial neural network and support vector machine. The binary choice model (logistic regresision) is other kind of popular model. It have similarities with the neural network.