JASIC Volume. 3, Issue 1 (2022)


Olanloye.D.O., Umezuruike.C, Emuoyibofarhe,O.N., Omideyi.D A.


Machine Learning Statistical Model Linear Regression Multiple Regression Mean Squared Error

Download Full-text (PDF)

... Download File [ 1.07 MB ]
Go Back

Prediction of the Price of Cement in South Western Part of Nigeria: A Machine Learning Approach

Abstract: Cement is one of the essential products that is needed in various construction works. It has been found very useful in the construction of houses, roads, bridges etc. In Nigeria the price of cement was quite stable in 1970’s and 1980’s. Recent market survey indicated that the price is on a very high side to the extent that it is almost becoming impossible for average Nigerians to have their own apartments where they can lay their heads after the daily activities. To the best of our knowledge, literature reveals that no attempt has been made to predict the price of cement in Nigeria, using Machine Learning (ML) models. Those that attempted it, mostly focused on statistical approach. In this research work, a machine learning model was used to predict the price using exponential regression model. Our method was compared with the existing ones and the performance is better. It produced the least Mean Squared Error (MSE) of 1.5455 and the percentage errors during the training were quite minimum. It was finally predicted that by the 2050, the price of 50kg bag of cement will be about ₦50,000 if government do not take necessary actions. This research work will be an eye opener to the government and other stakeholders.