The College of Administration and Economics at the University of Baghdad discussed , a master’s thesis in field of Statistics by the student (Yusra Jasim Suhail) and tagged with ( Estimating the nonparametric regression model for interval data with application ) , Under supervision of (Asst.prof.Dr. Saad Kadhem Hamza)
The use of parametric models requires the provision of many initial conditions for these models to be read correctly, which prompted researchers to search for models that are less complex than parametric models. These models are represented by nonparametric models. One of the methods used to estimate the nonparametric regression function is kernel estimation. Another problem we face is the presence of data in the form of intervals, as these data have internal variance that is not present in traditional data. Because of this variance, the development of new methods for analyzing this data was required. These methods are (N-W:CR, N-W:C, L-L-R:CR, CK+RL). These methods were applied and compared through a simulation method, in addition to using two models and three sample sizes (100, 200, 300) and different deviations. Observing the simulation results, it became clear that the best method is CK+RL. The best methods were also applied to real data for the percentage of oxygen in the blood (the independent variable) and heart rate (Dependent variable) for a sample of (206) observations of a group of patients (suffering from different types of illness) obtained from one of the hospitals in Baghdad (Martyr Dhari Al-Fayyadh General Hospital) for the period from 1/6/2024 to 30/7/2024. The results showed that the preference was for the CK+RL method.