TrORF: building trading areas around organizations based on machine learning techniques

Nowadays, making investments in trading areas around organizations is becoming increasingly pervasive since more and more organizations (e.g. universities) have been relocated in suburban-districts far from downtown. It generally requires a fairly long time for investors to make decisions and establish mature trading areas around organizations. Existing researches that focus on optimal location selection ignore the latent relationship between organizations and their surrounding business types. Therefore, this paper proposes a machine-learning based methodology of determining business types and trading areas, named TrORF, to help investors select business types and build trading areas around organizations. We implement our approach and evaluate it by using 13 real-world universities as our case studies. We compare the business type recommendation accuracy of the 13 real-world universities between our method and the other four existing methods. The experiment results indicate that our approach has strong ability to build the suitable business types around each university's trading area since its average recommendation accuracy is 98.23%, and it improves on other approaches by at least 13%.