Multivariate Analysis

Ultimately, data scientists must yield accurate predictions and classification to validate knowledge. Multivariate analysis enables data scientists to examine statistical relationships among data sets, including their aggregate effects on a dependent variable as captured by regression models . Of the various measurements of the statistical relationship between two or more variables, perhaps most common are those of correlation and regression coefficients. These measures are generally more complex than statistics that describe characteristics of a single set or variable. When we have many sets or fields of data, correlation and regression coefficients are computed using two sets at a time in a pair-wise or bi-variate manner. In that way, we determine statistical relationships between any two fields of data. In addition to correlation and regression, this chapter discusses the related topics of the coefficient of determination and statistics that are used in an ANOVA table.