Prediction of Customer Returns for an Online Shop
This is the submission for the final task for the Business Analytics and Data Science course from Prof. Dr. Lessmann at Humboldt University, Berlin. More information on the course can be found at https://github.com/Humboldt-WI/bads.
In this notebook I predict, if specific orders at an online retail shop will be returned by the customers. The goal is to predict such that an asymmetric cost function is minimized.
This is being achieved by data cleaning, feature selection, a custom-written scikit-learn datatransformer and logistic regression.