Solving unsupervised classification problems by new method

Unsupervised classification allows us to divide the dataset into several groups without knowing how the records should relate to each other. It is one of an interesting data mining topics that can be applied in many fields. A new method for solving this optimization problem is utilized. The method is based on the so-called Mesh Adaptive Direct Search method. This method does not explicitly use derivatives, and is particularly appropriate when functions are non-smooth, an important feature that has not been addressed in previous clustering studies. Results of computational experiments on real data sets present the robustness and advantage of the new method.