COMPARATIVE ALGORITHMS OF SUPPORT VECTOR REGRESSION MACHINE (SVRM) IN MODELING OF AIR PASSENGERS TRAFFIC
Olugbenga Ojo & Oluwatobi Ogunnusi , July 2020, Published: 2020-11-04
The rate of passengers’ traffic in Nigerian airports has increased exponentially in the last few years with significant contribution to economic growth. It is therefore important to know the future demand of air passengers’ traffic for the provision of proper air space resources. This paper focused on modeling air passengers traffic using Support Vector Regression Machine (SVRM), one of the Machine Learning (ML) techniques used in predicting future values of time spaced data. In achieving this, domestic flights data comprising of 132 months (2007-2018) collected from Nigeria Airspace Management Agency (NAMA) of the Muritala Mohammed International Airport was analyzed to achieve the set objectives. The preference evaluation of the adopted kernels in Python for data science was carried out on the basis of Root Mean Square Error. Empirical analysis of result showed that Linear Kernel function outperformed other kernel functions of Radial basis, Polynomial and Sigmoid respectively. Additional evidence was also shown to analyze the improved evidence of the linear kernel function from the prediction of the domestic air passengers with the use of the test data.