FRAC: A fast and robust algorithm for classification

In this paper, a hybrid approach was proposed which classified the test data with high accuracy and speed. At first, the data points were mapped to a range of 0 to 0.5 and, depending on the type of issue, they were divided to n regions. The accurate but slow algorithm was allocated to the regions close to (f(.)=0) decision function. For those regions which were far from decision function, a faster classification algorithm with less accuracy was allocated. The proposed method was tested on six public datasets and implementation of the results showed that the proposed method significantly reduced the classification time of tested data without reduction of its accuracy in comparison to the precision of the most accurate algorithm which was used. Also, it was demonstrated that, with the increase in the number of regions and allocation of the appropriate algorithm to it, time would reduce again.