프롭테크 업체 자동평가모형(AVM)의 적정성에 대한 연구

Abstract

In recent years, various services of proptech using big data from the public & private sectors and AI(Artificial Intelligence)technology based on big data have come into the world. The real estate environment is facing rapid changes with new services of protech based on big data. It is a very important time for the appraisal industry to actively respond and research. The most notable change in the appraisal industry is proptech’s AVM(Automated Valuation Model) value estimation service. Proptech’s services from big data is unavoidable and current trends. Therefore, it is desirable to promote a more efficient appraisal service by using these services as a mean of assistance in the appraisal industry. By comparing the results of proptech companies’ AVM value estimation with actual appraisal cases, it was analyzed whether the value estimation results of the automated valuation model were appropriate. The results of the AVM from the five proptech companies showed a gap of about 2.85% compared to the appraised value. However, the results of AVM value estimation on an individual lot indicated a difference about 240% and a gap up to 30% - 40% compared to the actual appraisal value, showing a large variance for each property. The results of AVM were appropriate only when the actual transaction price level in neighborhood was relatively consistent and frequent. Therefore, in order to increase the accuracy of AVM value estimation, it is necessary to develop a more sophisticated algorithm that can reflect the individuality of real estate.