Selective Inference for Sparse High-Order Interaction Models

no code implementations • ICML 2017 • Shinya Suzumura, Kazuya Nakagawa, Yuta Umezu, Koji Tsuda, Ichiro Takeuchi

Finding statistically significant high-order interactions in predictive modeling is important but challenging task because the possible number of high-order interactions is extremely large (e. g., $> 10^{17}$).

Drug Response Prediction feature selection +1

Selective Inference for Change Point Detection in Multi-dimensional Sequences

no code implementations • 1 Jun 2017 • Yuta Umezu, Ichiro Takeuchi

We study the problem of detecting change points (CPs) that are characterized by a subset of dimensions in a multi-dimensional sequence.

Change Point Detection Selection bias

Post Selection Inference with Kernels

no code implementations • 12 Oct 2016 • Makoto Yamada, Yuta Umezu, Kenji Fukumizu, Ichiro Takeuchi

We propose a novel kernel based post selection inference (PSI) algorithm, which can not only handle non-linearity in data but also structured output such as multi-dimensional and multi-label outputs.

General Classification Multi-class Classification