Modified Great Deluge for Medical Clustering Problems

Clustering problem is a type of classification under optimization problems, which is considered as a critical area of Data Mining. Medical clustering problem is a type of unsupervised learning in data mining. This work presents great deluge and modified great deluge algorithms for medical clustering problems. The structure of the modified great deluge (MGD) algorithm resembles a great deluge (GD) algorithm structure. The basic difference is that, in MGD the level is updated by a new level that is randomly selected from the list, whilst, in GD the level is initialized only once at the beginning of the search. Therefore, MGD has a better capability of escaping from a local optima compared to GD. Experimental results obtained by two way of minimal distance calculation tested over six benchmark medical datasets show that, MGD is able to produce significantly good quality solutions and outperform some instances of GD algorithm.