Sedef Hastalığının Tanı ve Tahmininde Yapay Zekâ Destekli Yeni Bir Yaklaşım

Early diagnosis and prediction of psoriasis is crucial to control disease progression, alleviate symptoms and reduce the risk of complications.Diagnosis in the early stages helps to determine the appropriate treatment plan and improve the patient's quality of life.The aim of this study is to enable early diagnosis of psoriasis.For this purpose, a hybrid architecture was created using a stacked auto-encoder, softmax classifier and Firefly Optimization Algorithm.With the created architecture, the architectural parameters of the stacked autocoder and softmax classifier hybrid structure aimed to be created for psoriasis diagnosis and all hyperparameters within the architecture were optimized.The model was implemented on the "Dermatology" dataset in the UCI data warehouse.In addition, machine learning methods such as K-Nearest Neighbor algorithm, Support Vector Machine and Decision Trees, which are frequently used in the literature, were also applied on the same dataset.The findings obtained from the experimental studies are presented in a controversial manner.The findings show that the proposed hybrid architecture achieves better results than other machine learning methods.At the same time, the model optimized and presented with the hybrid architecture can be used as an alternative method in patient decision support systems.