Friday, March 08, 2013 4:36 PM
What are the most frequently used DM models? There are many models in SSAS and I am wondering which are the ones I should focus on first. In the tutorial, they always focus on Decision Tree, Clusters, and Naive Bayes. Are these three the most frequently used? Of course, it all depends on the type of business one is trying to analyze, but which are the most demanding ones?
- Edited by BIAnalyst Friday, March 08, 2013 4:54 PM
Monday, March 11, 2013 7:11 AMAnswerer
I think most often people build classification models, that is to predict categorical columns.
Decision Tree algorithm has good accuracy and is easy to understand. Logistic Regression algorithm is as good but is much harder to understand.
Naive Bayes is used by "Analyze Key Influencers" in the Excel add-in, which is used more than others (again, easy to understand).
I think Time Series algorithm is used often too.
Clustering is used more often for data exploration.
Tatyana Yakushev [PredixionSoftware.com]
- Marked As Answer by BIAnalyst Monday, March 11, 2013 4:29 PM
Monday, March 11, 2013 4:30 PM
Thanks again, Tatyana!