Ranking Classi cation Algorithms with Dataset Selection: Using Accuracy and Time Results

Given that wide variety of available classiication algorithms exists, the selection of the right algorithm to use on a new problem is an important issue. In this paper we present zooming, that analyzes a given dataset and selects relevant (similar) datasets used in the past. This process is based on the \distance" calculated on the basis of several dataset characteristics. The accuracy and time results associated with the selected datasets are then processed to generate an advice in the form of a ranking, indicating which algorithms should be applied in which order. Here we propose the adjusted ratio of ratios ranking method. The generalization power of this ranking method is analyzed and the experimental results indicate that zooming leads to better results on average. The work presented can be seen as a rst step towards a system to provide advice on the utility of diierent solution strategies.