Chi-squared: A simpler evaluation function for multiple-instance learning
Amy McGovern, David D. Jensen · 2003
This paper introduces a new evaluation func-tion for solving the multiple instance problem. Our approach makes use of the main idea of di-verse density (Maron, 1998; Maron & Lozano-Pérez, 1998) but finds the best concept using the chi-square statistic. This approach is sim-pler than diverse density and allows us to search more extensively by using properties of the con-tingency table to prune in a guaranteed man-ner. We demonstrate that this approach solves the multiple-instance problem as well as or better than diverse density and that the pruning mecha-nism allows chi-squared to identify the best con-cepts more quickly. 1.