The effect of google drive distance and duration in residential property in Sydney, Australia

Publication Type:
Conference Proceeding
Citation:
Uncertainty Modelling in Knowledge Engineering and Decision Making - Proceedings of the 12th International FLINS Conference, FLINS 2016, 2016, pp. 646 - 655
Issue Date:
2016-01-01
Full metadata record
Files in This Item:
Filename Description Size
Finalised Paper 189. Mehrdad.pdfAccepted Manuscript version393.77 kB
Adobe PDF
© 2016 by World Scientific Publishing Co. Pte. Ltd. Predicting the market value of a residential property accurately without inspection by professional valuer could be beneficial for vary of organization and people. Building an Automated Valuation Model could be beneficial if it will be accurate adequately. This paper examined 47 machine learning models (linear and non-linear). These models are fitted on 1967 records of units from 19 suburbs of Sydney, Australia. The main aim of this paper is to compare the performance of these techniques using this data set and investigate the effect of spatial information on valuation accuracy. The results demonstrated that tree models named eXtreme Gradient Boosting Linear, eXtreme Gradient Boosting Tree and Random Forest respectively have best performance among other techniques and spatial information such drive distance and duration to CBD increase the predictive model performance significantly.
Please use this identifier to cite or link to this item: