Do you want to know what will happen in the future? To gain true predictive insight, skip the tea leaves and look toward your data. SAS instructor Jeff Thompson is a high-energy data mining expert who will be demonstrating how to gain predictive insight from your data in his new course, Mass-Scale Predictive Modeling using SAS Factory Miner, which will make its debut after the Analytics 2015 conference in Las Vegas.
I recently caught-up with Jeff to ask him a few questions about his new course.
- Why should people get excited about this course?
People should be excited about the course because it is brand new! The product students will learn is brand new! And there really is nothing else in the market today that has all the capabilities that SAS Factory Miner has when it comes to automated, segmented modeling.
- Who would get the most out of attending this course?
Certainly customers who have recently acquired SAS Factory Miner. That product was just released this past July. But even modelers without the product could benefit. Segments exist in many data sets and the course focuses on predictive modeling based on segments in data.
- How long have you been working with SAS Factory Miner?
I recall first seeing Factory Miner about a year or so ago when it was still in the early stages of development. I really started to work with this new product to learn it and begin to develop the course in the spring of 2015. This was still several months before the product was released.
- What was your favorite part about developing this course?
Working with brand new and cutting edge technology. It is very exciting to be working on a course for the initial release of a new product. Also, it has been great to work with my colleague, Peter Christie, on the course’s development. Peter comes from more of a business background whereas mine is pure statistics. I have learned quite a bit from him during the process and the different skill sets we each bring to the table have really come together to create a great course.
- Is there any background required before taking this course?
Yes, we suggest students be familiar with the basics of predictive modeling, and although it is not used in the course, some knowledge of Enterprise Miner would be helpful. Factory Miner builds models using modern machine learning algorithms including, but not limited to, random forests, support vector machines, and neural networks. We do not cover the details of these modeling methods so having some basic knowledge of them prior to taking the course is beneficial. Two courses where this background can be picked up are Applied Analytics using SAS Enterprise Miner and Strategies and Concepts for Data Scientists and Business Analytics.
- What kind of business problems can customers solve after taking this course?
This is an important question to answer. Recently customers have realized that building only a single or small number of models for their customer database is not sufficient. Further, the process of manually constructing different models for many segments is time consuming and expensive. We can get closer to one-to-one marketing if we can build models for the many segments that exist within the enterprise data and having a fully automated way to do so saves valuable time and money. Segments may be based on geography, customer status such as “VIP” levels, or any of several socio-demographic variables. Factory Miner allows segments to be constructed by combining levels of up to three segment variables and a modeling tournament is ran on all levels of the resulting segment profile. But the automated process is “white-box” where modelers can modify default model templates or create their own.
Get registered to learn more from Jeff at his two-day course in Las Vegas.