Machine Learning-Based Assessment of Psoriasis Using Real-World Clinical Data from Egyptian Population

Psoriasis is a chronic autoimmune disease that impacts millions of people globally, with varying degrees of severity, and in turn, affects their quality of life. In the current work, we used a clinical dataset of Egyptian psoriasis patients to build machine learning models for two major tasks: differentiating psoriasis patients from healthy controls and predicting disease severity. The dataset consists of demographics, clinical evaluations, genetic factors, and treatment response data. For the purpose of classification, we compared the outputs of a series of models, including Logistic Regression (LR), Random Forest (RF), Extra Trees, Gradient Boosting, Extreme Gradient Boosting (XGBoost), Light Gradient-Boosting Machine (LightGBM), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Neural Networks, and found the best accuracy rate at 87% through the Random Forest model. The prediction of severity was approached using the following models: Gradient Boosting Classifier ( GBC), R andom F orest, e Xtreme G radient Boosting (XGB) Classifier, S upport Vector M achine ( SVM), L ogistic Regression (LR), and Light Gradient-Boosting Machine (LGBM) Classifier, in which the highest performance was attained by the XGB Classifier ( 86%), p articularly i n p redicting s evere cases. The research supports the potential of using machine learning in enhancing psoriasis treatment by classifying patients effectively and forecasting disease severity, providing valuable information for precision medicine in dermatology.