Title: Quantile regression approach to conditional mode estimation
Authors: Kengo Kato - Cornell University (United States)
Satoshi Hara - Osaka University (Japan)
Hirofumi Ohta - The University of Tokyo (Japan) [presenting]
Abstract: The estimation of the conditional mode of an outcome variable given regressors is considered. To this end, we propose and analyze a computationally scalable estimator derived from a linear quantile regression model and develop non-standard asymptotic distributional theory for the estimator. Specifically, we find that the limiting distribution is a scale transformation of Chernoff's distribution, i.e., the distribution of a maximizer of a two-sided Brownian motion with a negative quadratic drift, despite the presence of regressors. In addition, we consider analytical and subsampling-based confidence intervals for the proposed estimator. We also conduct Monte Carlo simulations to assess the finite sample performance of the proposed estimator together with the analytical and subsampling confidence intervals.