Prediction Model of Commitment of New Self-Employed Omanis to Insurance Systems.

The Omanis self-employed insurance system is one of the four insurance systems for private-sector employees that the Public Authority for Social Insurance (PASI) supervises. The system suffers from some problems related to the contributors' payment commitment. The reason is that some factors affect their commitment after joining the system. The payment defaulting presents a challenge for PASI as it becomes an obstacle to realizing the benefits. The study aims to identify the most critical factors that affect contributors’ commitment to this system. The data mining and machine learning methods are used, such as Logistic Regression Classifier, Decision Tree Classifier, Support Vector Machine Classifier, and Multilayers Perceptron Classifier, to identify these factors and classify committed contributors. A sample was obtained from 11773 contributors registered in PASI data with 24 features. This data has also been reprocessed to get rid of noisy and irrelevant data. The study concluded that the most critical factors that affect contributors' commitment to this system are personal factors, commercial factors, and economic factors. Besides, the results of the recommended model were 0.94, which is the Logistic Regression Classifier. The study recommended using machine learning as the primary approach in the social insurance field in Oman. It is more accurate and efficient than traditional statistical methods to help decision-makers on scientific foundations.