An effectiveness of machine learning models for estimate the financial cost of assistive services to disability care in the Kingdom of Saudi Arabia.

As per the World Health Organization (WHO), the justifications of people with disabilities globally are restricted by physical and social barriers that exclude their full contribution to society. Constructed environment barriers can limit the availability of transportation, employment, goods and services, healthcare, and overall independent drive. The government of Saudi Arabia has applied programs and policies to enhance the quality of life for people with disabilities, including education, healthcare, and employment chances. Furthermore, they also take action to progress a few social guards that endorse public involvement and income-support plans for individuals with disabilities, besides efforts to uphold the cultural, social, political, and economic environment for accurate plans. Therefore, this study presents the Effectiveness of Machine Learning Models for estimating the Financial Cost of Assistive Services to Disability Care (EMLM-EFCASDC) technique in the KSA. The presented EMLM-EFCASDC technique mainly aims to develop a data-driven model that accurately predicts the cost of assistive services in disability care across the KSA. At first, the EMLM-EFCASDC approach utilizes Z-score normalization to preprocess the input data, ensuring that data variability is minimized for improved model accuracy. Next, an ensemble of machine learning (ML) models comprises three classifiers such as hybrid kernel extreme learning machine (HKELM), extreme gradient boosting (XGBoost), and support vector regression (SVR) for predicting the financial cost. Eventually, the modified pelican optimization algorithm (MPOA) is utilized to fine-tune the optimal hyperparameter of ensemble model parameters to achieve high predictive performance. An extensive range of simulation analyses are employed to ensure the enhanced performance of the EMLM-EFCASDC technique. The performance validation of the EMLM-EFCASDC method portrayed the least RMSLE value of 0.1154 on existing approaches in terms of diverse evaluation measures.