Title :
Online local linear classification
Author :
Wang, Jiacheng ; Trapeznikov, Kirill ; Saligrama, Venkatesh
Author_Institution :
Dept. of Electr. & Comput. Eng., Boston Univ., Boston, MA, USA
Abstract :
We present a novel convex formulation to learning binary, 2-region local linear classifiers. From this convex formulation, we formulate an online optimization scheme using stochastic gradient descent that allows for efficient training using streaming training data. We demonstrate the fast convergence and accurate classification on the canonical XOR dataset.
Keywords :
convex programming; data handling; gradient methods; learning (artificial intelligence); pattern classification; training; canonical XOR dataset; convex formulation; learning binary classifier; online local linear classification; online optimization; stochastic gradient descent method; streaming training data; two-region local linear classifier; Fasteners; Support vector machine classification;
Conference_Titel :
Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), 2013 IEEE 5th International Workshop on
Conference_Location :
St. Martin
Print_ISBN :
978-1-4673-3144-9
DOI :
10.1109/CAMSAP.2013.6714035
