Classifier selection process implies mastering a lot of background information on the dataset, the model and the algorithms in question. We suggest that a recommender system can reduce this effort by registering background information and the knowledge of the expert. In this study we propose such a system and take a first look on how it can be done. We compare various classifiers against different datasets and then come up with the most appropriate classifier for a particular dataset based on its unique characteristic.
Ahsan, Mujtaba and Ngo-Ye, Lin, "A Conceptual Model of Recommender System for Algorithm Selection" (2005). AMCIS 2005 Proceedings. 122.