On the Partial Least Square Regression Modeling to Collinear Regressors: Contribution of Transportation Sector to Nigeria Economic Growth

Ogunnusi, O.N and Ojo, G and Sikiru, O. A. (2021) On the Partial Least Square Regression Modeling to Collinear Regressors: Contribution of Transportation Sector to Nigeria Economic Growth. IJISET - International Journal of Innovative Science, Engineering & Technology, 8 (4). pp. 231-241. ISSN 2348 – 7968

Text IJISET_V8_I04_24.pdf Download (426kB)

Abstract

This study dwelt on the application of partial least square regression (plsr) modeling using the contribution of transportation sector to economic growth data. Twenty-nine (29) years data covering 1981 to 2019, extracted from Central Bank of Nigeria statistical bulletin which consist the contribution of road, air, water, rail, transport service, post and couriers’ services was proxied with economic growth to confirm its pattern of contribution of the predictor variables. Ordinary Least Square Regression model was first fitted to the data to confirm if there exists multicollinearity in the set of predictors. Result of the Variance Inflation Factors (VIF) technique adopted indicated severe multicollinearity among the set of predictors of road, rail, transport service, post and courier services with associated coefficients of 33.0751, 7.3543, 21.2836 and 62.0324, violating the assumption of predictors independence. The partial least square model fitted using kernel function of the cross-validation technique indicated maximum of three components explaining 99.99% variance of the predictors and 99.24% variance of economic growth. Comparative analysis of the out of sample predictions in R for data science was carried out using RMSE evaluation technique as this indicated that PLSR is better fitted to data with multicollinearity effect as recorded from the lower RMSE of predictions.

Item Type:	Article
Subjects:	Q Science > Q Science (General) Q Science > QA Mathematics
Divisions:	Faculty of Engineering, Science and Mathematics > School of Mathematics
Depositing User:	Mr. Bolanle Yisau I.
Date Deposited:	17 Feb 2022 19:44
Last Modified:	17 Feb 2022 19:44
URI:	http://eprints.federalpolyilaro.edu.ng/id/eprint/1939

Actions (login required)

View Item