Combining Heterogeneous Classifiers for Relational Databases

TL;DR

This paper introduces a hierarchical meta-classification method for relational databases that efficiently combines heterogeneous classifiers across multiple tables, reducing computation time while maintaining accuracy.

Contribution

It presents a novel recursive aggregation technique for heterogeneous classifiers in relational databases, addressing semantic and computational challenges.

Findings

Significant reduction in classification time

No loss in prediction accuracy

Effective on diverse benchmark datasets

Abstract

Most enterprise data is distributed in multiple relational databases with expert-designed schema. Using traditional single-table machine learning techniques over such data not only incur a computational penalty for converting to a 'flat' form (mega-join), even the human-specified semantic information present in the relations is lost. In this paper, we present a practical, two-phase hierarchical meta-classification algorithm for relational databases with a semantic divide and conquer approach. We propose a recursive, prediction aggregation technique over heterogeneous classifiers applied on individual database tables. The proposed algorithm was evaluated on three diverse datasets, namely TPCH, PKDD and UCI benchmarks and showed considerable reduction in classification time without any loss of prediction accuracy.