Improving Data Cleaning Using Discrete Optimization

Kenneth Smith, Sharlee Climer

One of the most important processing steps in any analysis pipeline is handling missing data. Traditional approaches simply delete any sample or feature with missing elements. Rece…