Week 1
I met with Dr. Dutta for the first time this week.
Prior to our meeting, I prepared by reading background material on Classification from 'Machine Learning' by Tom Mitchell. We agreed to meet on Mondays at 11:00, for an hour. Dr. Dutta gave a basic description of the project: constructing Probability Estimation Trees to solve Ranking. She outlined her expectations for the course of the project and assigned readings in chapter 6 of
Data Mining: Concepts and Techniques and 'Learning to Rank using Classification and Gradient Boosting' in order to differentiate between the machine learning problems of Classification and Ranking. She also assigned 'Tree Induction for Probability Based Ranking' as an introduction to Probability Estimation Trees, and to download and install WEKA, a machine learning software.
Downloading WEKA was simple, but I must have done something wrong, because it shut itself down every time I worked with it for more than a minute. It turned out to be because I had multiple installations on my computer &mdash I'm still not sure how. Once I had successfully dealt with this, WEKA's basic features turned out to be fairly simple to understand, although its more powerful tools are still incomprehensible to me. I wasn't able to find a tutorial simple enough for me to understand, but I hope to learn more about it as I work with it more.