Armin Shmilovici Leib

Senior Academic

A Working Set Algorithm for Accelerating the Relevance Vector Machine

David Ben-Shimon, Armin Shmilovici
The runtime complexity of the Relevance Vector Machine (RVM) is which makes it too expensive for moderately sized machine learning problems. We propose a working set algorithm which reduces the runtime complexity to O(N2). Experiments verify the viability of the method for classification benchmark problemsץ
Publication language English
Pages 56-59
Publication status Published - 2006

Keywords

Machine Learning
Data Mining
The Relevance Vector Machine
Working Set