R bloggers

Exploring lime on the house prices dataset

FavoriteLoadingAdd to favorites

Pretty recently I found a paper with the title “Why Should I Trust You?”: Explaining the Predictions of Any Classifier. The topic of interpretability is very important in the times of complex machine learning models and it’s also related to my PhD topic (reliability of machine learning models). Therefore I wanted to play around with the method a little bit. The method that was introduced in the paper is called LIME (Local Interpretable Model-Agnostic Explanations) and comes with a python package. Luckily for me, someone already ported it to R (thomasp85/lime). In this post I will show how I used LIME on regression models. What is LIME? LIME is a tool for explaining what a complex (often called black-box) machine learning model does. This is achieved by learning simple (e.g. linear regression) models on perturbed input data to figure out…
Original Post: Exploring lime on the house prices dataset