In this paper, we propose a new nonparallel support vector classifier, named ν-improved nonparallel support vector machine (ν-IMNPSVM), for solving binary classification problems. The model is adjusted on the nonparallel support vector classifier. In order to control the number of support vectors effectively, the value of C is replaced by the quantity ν. At the same time, ε-band is kept as small as possible. The hyperplane is constructed in the region where the data points are densely distributed, which plays a great role in the processing of unbalanced data sets. ν-IMNPSVM inherits the advantages of the previous model and improves the classification effect to some extent. The feasibility and effectiveness of the new model are showed by the validation results of a large number of data sets.
AMS Subject Classification: 90C30