Factored built an easy-to-use pipeline so that simple configuration files could be used to generate ensemble and boosting models. This delivered automatic feature interpretability and prediction capabilities.
The Challenge:
To automatically generate models for a telco company by using just input metadata and a configuration file, and extract the feature importance with directionality.
The Solution:
Implemented a pipeline that is able to train tree-based and boosting models for regression and classification problems, allowing the use of hyperparameter optimization and random/custom splits as well as the execution of different methods for feature importance extraction.
The Outcome:
An easy-to-use pipeline that receives one input and one modeling configuration file and automatically produces a fitted model with its evaluation and a new file containing a summary of the feature interpretability.
Tech Stack & Skills:
Treeinterpreter, SHAP, Scikit-learn, LightGBM, SciPy, Hyperopt, AWS EC2, XLWings.
Model Interpretability, Ensemble and Boosting Methods, Model Evaluation, Excel-Python Integration.
Results
Stuff
Stuff
Stuff