Generalized Low Rank Models

Across business and research, analysts seek to understand large collections of data organized as a table with numeric, Boolean, and categorical values. Many entries in the table may be noisy or even missing altogether. Low rank models facilitate understanding of tabular data by producing a condensed vector representation for every row and column in the data set. These representations can then be compared, clustered, plotted, and used in subsequent analysis. In this presentation, we will describe what a low rank model is and demonstrate how to build them in H2O. Through examples, we'll see how to fit low rank models to numeric and categorical data sets with missing values, and how to use these models to identify important features and make better predictions.