Classification, as a machine learning technique, can be applied for problem solving in almost every aspect of our lives. That is the reason why today there are lot of specialized classification systems and tools, based on different methods, whose purpose is to solve a specific problem. In this paper, we propose a model of a generic classification system based on a multiple kernel data fusion. The model describes a system that does a data preprocessing, intelligent choice of a classification method, proposes a solution and automatically tunes the parameters. This model is also parallel and scalable and it integrates multiple heterogeneous data sources. The main goal is to create a simple system that will focus the user on a method that he/she would further tune to obtain the final solution of a given problem.
models of classification systems, generic classification, multiple kernel data fusion