Decision tree-based methodology to select a proper approach for wart treatment

Abstract

The human papillomaviruses (HPVs) can be responsible for various types of benign tumors called warts. Although warts can grow on all parts of the human body, common warts and plantar warts (as the most prevalent warts) grow principally on the hands and feet soles, respectively. Different treatment approaches such as cryotherapy and immunotherapy can be used to conquer the disease. However, the best healing method should be selected based on the patient circumstances. This study employs the classification and regression tree (CART) algorithm to develop accurate predictive models capable of analyzing the response of patients having common and/or plantar warts to the cryotherapy and/or immunotherapy methods. To develop a CART classifier for the cryotherapy method, independent parameters including the age and gender of patient, number of warts, type of wart, surface area of warts, and the time elapsed before treatment are used. In the case of immunotherapy, in addition to the above-mentioned variables, the induration diameter of the initial test is also considered. The error analysis reveals that the implemented CART models provide the highest achievable accuracy for the application of interest. Moreover, the proposed decision tree-based models are simple to use and more reliable, in contrast to the literature models that are mainly originated from the fuzzy rule-based method. Hence, the models introduced in this study can assist both patients and physicians save cost/time and improve the quality of healing operation.

Keywords: Classification; Cryotherapy; Decision tree; Error analysis; Immunotherapy; Warts.