Automated Feature Engineering Applied to Causality
published: Oct. 6, 2014, recorded: December 2013, views: 2350
Report a problem or upload filesIf you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
This cause-effect pairs challenge was motivated by the contrast between the costs to per- forming controlled experiments in order to determine causality and the abundance of observational data. Our goal was to provide a value representing our confidence of causality determined by the observation data which would help identify the most promising variables for experimental verification of their causal relationship. A novel approach was created focusing on feature engineering that requires minimal human intervention. By applying standard machine learning algorithms to the pairs of points, almost 9000 features were created by computing the goodness of fit of these algo- rithms in various ways. This approach was successful enough to attain the highest score in the competition’s private leaderboard. Additionally, alternatives and their explanations of why they weren’t used as well as possible improvements which could greatly improve accuracy were outlined.
Link this pageWould you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !