A Classification via Clustering Approach for Enhancing the Prediction Accuracy of Erythemato-Squamous (Dermatology) Diseases

Identification of various erythemato-squamous diseases is a tangible problem in dermatology. Since, they all share the clinical features of erythema and scaling, only with very slight variances. The diseases in this group include seboreic dermatitis, cronic dermatitis, lichen planus, pityriasis rubra pilaris, pityriasis rosea and psoriasis. Generally a biopsy is required for the analysis but unfortunately these diseases share similar histopathological features as well. Another problem in the differential diagnosis is that a disease show the same feature of another disease at the initial stage and can only be characterized at the subsequent stages. Patients are clinically evaluated with 12 features at the beginning stage and then the skin samples are considered for the evaluation of 22 histopathological features. The data of the histopathological features are determined by an analysis of the samples under a microscope. This paper presents a predictive model of dermatological diseases by thoroughly analyzing both the clinical and histopathological features.