Conclusion

Abstract

In this book, we presented several recent advances of search techniques in intelligent classification systems with the medium-sized databases. We particularly focused on the small sample size problem, when only few instances are available for each class. We reduced the classification task to the hypothesis testing for homogeneity of features. Based on this approach the methodology of segment homogeneity testing was introduced. In this chapter we discuss potential applications of this methodology to increase the computing efficiency of existing intelligent systems with the large number of classes.