A New Stacking Method For Winning A Gold Medal In Zillow Prize Competition: The Triad Method

The Zillow Prize challenge on Kaggle was one of the most popular competitions with more than 4,000 participants. Our team at Prizmi were among the participants and won two medals. To win those competitions, we invented and implemented a novel approach to combine the predictions of several models. We are happy to announce that Leaner now supports this method. In this article, we introduce the Triad method and show how it can be done using Leaner without writing any code.

Build Production-Ready Deep Learning Models Without Writing Code

We are happy to introduce two new engines in Learner: DeepClassifier and DeepRegressor. Using these engines you can build production-level deep learning models on your tabular data to solve business problems. In this post, we'll go over some background on using Deep Learning models on tabular data . We'll then show how you can build such models in Learner without writing code.