From the course: Predictive Analytics Essential Training: Data Mining

Unlock the full course today

Join today to access over 23,200 courses taught by industry experts.

Addressing missing data

Addressing missing data

- [Instructor] I was visiting a client site some years ago and they asked me to do an audit of their work. It was a fun project because they did excellent work, I was just looking for opportunities for improvement. I noticed that the software was consistently choosing algorithms that had a particular trait. The winning algorithms were automatically handling missing data. The other algorithms, including some of the best, were never chosen, and I suddenly realized they didn't have a missing data strategy. They were counting on the software to do it for them. Now, entire books have been dedicated to missing data. Nonetheless, there are a couple of general observations that can be made. First, you will have missing data every single time. How can I be so sure? Well, it's not just about data cleanliness, of course, if the data's not clean, you've got problems, but everyone will have situations that are not applicable, which is…

Contents