Adversarial Validation to Select Validation Data for Evaluating Performance in E-commerce Purchase Intent Prediction
In recent years, there has been a tremendous growth in research on machine learning techniques such as deep learning. However, regarding real-world problems, practitioners still face many challenges on how to prepare data and how to validate machine learning models for specific problems. This paper describes a method for validation data selection to apply adversarial validation methodology, creating a classifier for training and test data. This approach evaluates machine learning models, taking the differences between training data and test data into account. We applied it to the purchase intent predicting task in the Coveo Data Challenge: 2021 SIGIR Workshop on eCommerce. This approach guided us in the challenging task to achieve the third place in the leaderboard. Our codes are available at https://github.com/upura/sigir-ecom-2021/.