Relationship discovery in business analytics
A subset of (k−1)-dimensional tables are received, wherein k is greater than 1. A set of k-dimensional tables is created by combining each of the (k−1)-dimensional tables with a non-included dimension corresponding to a 1-dimensional table. Significance of interaction and interaction effect size is computed for the created set of k-dimensional tables to determine dimension and measure interactions.
1. A method comprising:
receiving, using a processor of a computer, a subset of (k−1)-dimensional tables, wherein k is greater than 1, wherein the subset of (k−1)-dimensional tables are selected based on goodness of fit values exceeding a first threshold, wherein the goodness of fit values are statistics that evaluate model capability to describe data;
creating a set of k-dimensional tables by combining each of the (k−1)-dimensional tables with a non-included dimension corresponding to a 1-dimensional table;
outputting a subset of the k-dimensional tables selected from the created set of k-dimensional tables for use in creating (k+1)-dimensional tables, wherein the subset of the k-dimensional tables are selected based on goodness of fit values exceeding a second threshold, and wherein the created k-dimensional tables are sorted according to the goodness of fit values; and
computing significance of interaction and interaction effect size for the created set of k-dimensional tables to determine dimension and measure interactions.
2. The method of claim 1 , further comprising:
sorting the created set of k-dimensional tables that have significant interaction effect according to the interaction effect size.
3. The method of claim 2 , further comprising:
using a subset of the sorted k-dimensional tables that have an interaction effect size exceeding a third threshold to generate one or more reports.
4. The method of claim 1 , wherein the significance of interaction and the interaction effect size are computed for the set of created k-dimensional tables using interaction indices in terms of summaries derived from aggregated input statistics that aggregate measure values.
5. The method of claim 1 , further comprising:
receiving aggregated input statistics for 1-dimensional tables;
computing the goodness of fit values for each of the 1-dimensional tables; and
sorting the 1-dimensional tables according to the computed goodness of fit values.
6. The method of claim 1 , wherein the subset of (k−1)-dimensional tables comprise top (k−1) dimensional tables selected based on the goodness of fit values.
7. The method of claim 1 , further comprising:
computing the goodness of fit values using an Analysis of Variance (ANOVA) model.
8. The method of claim 1 , wherein a Software as a Service (SaaS) is provided to perform the method.