Handling missing data

In many cases, several data points may have values missing from certain features. If you're looking at Yes/No responses to survey questions, several participants may have accidentally or purposefully skipped a given question. If you're looking at time series data, your measurement tool may have had an error for a given period or measurement. If you're looking at e-commerce shopping habits, some features may not be relevant to a user, for instance last login date for users that shop as an anonymous guest. The inpidual situation and scenario, as well as your algorithm's tolerance for missing data, determines the approach you must take to remediate missing data.