Machine learning to assess business-to-business relationships

The document discusses using machine learning to develop a relationship health score that assesses the health of business-to-business relationships. It details extracting features from a company's database covering subscription data, relationship details, customer service notes, communications, and using them to build a random forest model to predict the probability of subscription renewal. This probability is then transformed into a relationship health score. Feature engineering improved the model's accuracy dramatically, decreasing its log loss by 82%.