The presentation below, “Survey of Available Machine Learning Frameworks,” is provided by Brendan Herger of CapitalOne as part of the H2O World 2015 conference. Learning a new modeling framework is time consuming, and doesn’t always pay off. However, as more feature engineering and modeling frameworks become available, its difficult not to leverage their abilities. Only interested in frameworks available in R? Need a large selection of clustering and regression algorithms? Limiting your train data set because your framework is bursting at the seams? The talk covers an in depth overview of the strengths, weaknesses and design logic of the top feature engineering and modeling frameworks available, and which of these frameworks justify pushing through the learning curve.
Here are the slides that accompany the presentation:
Sign up for the free insideBIGDATA newsletter.