Tin Kam Ho (Chinese: 何天琴) is a computer scientist at IBM Research with contributions to machine learning, data mining, and classification. Ho is noted for introducing random decision forests in 1995, and for her pioneering work in ensemble learning and data complexity analysis.

Following the principles of stochastic modeling, we propose a method to construct tree-based classifiers whose capacity can be arbitrarily expanded for increases in accuracy for both training and unseen data. The essence of the method is to build multiple trees in randomly selected subspaces of the feature space.

Advances in Pattern Recognition: Joint IAPR International Workshops SSPR'98 … I Guyon, K Bennett, G Cawley, HJ Escalante, S Escalera, TK Ho, N Macià, ... IEEE Transactions on Pattern Analysis and...

Tin Kam Ho AT&T Bell Laboratories 600 Mountain Avenue, 2C-548C Murray Hill, Abstract Decision trees are attractive classifiers due to their high execution speed. But trees derived with traditional methods often cannot be grown to arbitrary complexity for possible loss of generalization accuracy on unseen data.

Tin Kam Ho (S'89–M'91–SM'02–F'06) received the Ph.D. degree in computer science from SUNY at Buffalo, Buffalo, NY, USA, in 1992. She is a Research Staff Member of IBM Watson Group. Formerly, she was with Bell Labs Research, where she led the Statistics and Learning Research Department in 2008–2014.

ABSTRACT: Machine Learning has undergone a tremendous progress, which is evolutionary over the last decade. It is widely used to make predictions that lead to the most valuable decisions.

Tin Kam Ho Senior AI Scientist, Watson Health, International Business Machines. Joined ; June 2017

On classification, I explored methods for multiple classifier systems, random decision forests, and more recently, data complexity analysis. To facilitate these, I also explore methods and tools for interactive data visualization and analysis.

Tin Kam Ho Ester Bernadó-Mansilla We study the domain of competence of a set of popular classifiers, by means of a methodology that relates the classifier’s behavior to problem complexity.