Determinants of Urban Land Lease Price Using Artificial Neural Network and Hedonic Regression Model: Case of Addis Ababa, Ethiopia

Abstract

The objective of the current Ethiopian lease proclamation is to ensure efficiency and transparency in land delivery and to enhance the revenue generated to the state through the transfer of land by tender. However, there are various challenges in relation to access to urban land in Ethiopia that adversely affects the business environment. Thus, this study is aimed at identifying the determinants of urban land lease price using artificial neural networks (ANNs) and hedonic regression methods. The data about land lease price have been collected from the Addis Ababa city land transfer office. After coding and editing, the data have been processed in statistical software. Based on a significance threshold larger than 0.05, it was discovered that the three variables in the hedonic regression model (HRM) have no effect on the price of urban land leases. In the instance of the ANN, all variables have an effect on the price of Addis Ababa’s urban land leasing. When the results of the two methods are combined, all of the variables have an impact on the determination of Addis Ababa’s urban land leasing price. The ANN outperforms the HRM in terms of model prediction.