The study pertaining to the available medical databases which are of huge size involve the consideration of all the attributes related to the problem. This resulted in large amount of computational time and accordingly there was a drastic reduction in the computational speed also. In order to minimize these factors, the present study was undertaken which enabled to get optimal results by using neural classifier techniques. In this paper, the effectiveness of various attributes and classifiers in the cytological diagnosis of WBCD breast cancer dataset were compared. Here, the most effective attributes were identified and it was found that these attributes describe at least one of the important nuclei characteristics of the morphological and textual features namely size, shape and texture of a cancerous cell. Further, applying all these attributes together to the classifiers, it was found that there was a significant increase in their performance which resulted in optimal computational speed and time. Overall, it was found that support vector machines could give accuracy as high as 97.37%.