Identifying Contributors That Explain Differences Between Subsets of a Data Set

摘要

Methods for analyzing and rendering business intelligence data allow for efficient scalability as datasets grow in size. Human intervention is minimized by augmented decision making ability in selecting what aspects of large datasets should be focused on to drive key business outcomes. Variable value combinations that are predominant drivers of key observations are automatically determined from several competing variable value combinations. The identified variable value combinations can then be then used to predict future trends underlying the business intelligence data. In another embodiment, an observed outcome is decomposed into multiple contributing drivers and the impact of each of the contributing drivers can be analyzed and numerically quantified—as a static snapshot or as a time-varying evolution. Similarly, differences in observations between two groups can be decomposed into multiple contributing sub-groups for each of the groups and pairwise differences among sub-groups can be quantified and analyzed.

主权项

1. A method for analyzing differences in an outcome between a subset A from a data set for a process and a subset B from the same data set, the method comprising a computer system automatically performing the following:
processing a data set containing observations of the process, the observations expressed as values for a plurality of variables and for the outcome, wherein processing the data set determines behaviors for different variable combinations with respect to the outcome, the variable combinations defined by values for one or more of the variables, the subset A defined as those observations for which one or more test variables take first values and the subset B defined as those observations for which the test variables take different second values; for pairs of a first variable combination and a second variable combination, wherein the first and second variable combinations are the same except that the test variables take the first values in the first variable combination and take the second values in the second variable combination, estimating contributions of the pair to differences in the outcome between subsets A and B, based on differences in the behaviors of the pair and also based on differences in populations of the pair; and reporting differences in the outcome between the subsets A and B based on the estimated contributions for the variable combinations.