Auto-Indexing Selection Technique in Databases under Space Usage Constraint Using FP-Growth and Dynamic Programming

Abstract

This paper presents an auto-indexing selection technique to improve query processing time under indices' space usage limitation. The technique is composed of two steps: candidate extraction and index selection. In the candidate extraction step, data mining technique called FP-growth is used to find relationships among attributes. Then candidate indices are extracted. By giving space usage constraint on the index selection step, the dynamic programming approach combined with memory functions technique is used to find the most valuable subset of the indices among the candidates.