The Application of Feature Selection under Supervised Learning in Liquid Recognition

The problem needed to be solved firstly in liquid identification based on Droplet Fingerprint is feature selection. Feature selection is how to minimize the dimension of feature under the premise of not reducing the classification accuracy. According to the DB index classification criteria and floating search algorithm, this paper proposes an algorithm of feature selection suitable for supervised learning and performs the algorithm in Droplet Fingerprint database. Experimental results show that the algorithm performs better and greatly improves the efficiency of feature selection. (C) 2011 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of [CEIS 2011]