# debut-forecasting-models
Debut Forecasting Modeling 

## Problem: 
Forecast the performance of track. These are broken down into pre-release and post-release. In the pre-release case, we have to do this without any prior performance data about the new track. In the post-release case, we have some performance data about and we are interested in forecasting the decay of the model.

## What is included:
- feature_preprocessing - pipelines for feature engineering 
- model_evaluation - pipelines for model evaluation and logging of model perforamcen on mlflow
- models - for modeling pipleines which includes the feature engineering and modeling pipelines
- automl (coming soon) - this bit will finetune models to find the best hypeparameters