CMDA 4654 Study Guide - Final Guide: Marginal Distribution, Point Cloud, Conditional Variance

• School: Virginia Tech
• Department: Computational Modeling and Data Analytics
• Course: CMDA 4654
• Professor: R B Gramacy

Document Summary

In r, factors are considered categorical variables where such a factor can take on levels. Typically, a covariate describes an input or x-variable, but sometimes, people also refer to the response, y, as a covariate more generally. Regression is really about modeling the conditional distribution of y given x. If the data is comprised of (xi, yi) pairs, then we can use these to learn about the conditional distribution. A conditional distribution can be obtained by slicing the x y point cloud. In the example in the slides, the marginal distribution would show a boxplot of the entire data. The marginal distribution can also show the variability within groups. On the other hand, the conditional distribution would slice the data based on a particular. X-variable (e. g. , say ranges of square footage by thousand). In the example in the slides, the conditional distribution can give a point forecast (the median) and a prediction interval for a particular range of square footage.