►A NOVEL SALES PREDICTION APPROACH: Competing vs Supporting Networks
Kun Chen, Dejun Xie and Huaiqing Wang
10.52283/NSWRCA.AJBMR.20150410A03
ABSTRACT
Modelling and predicting online sales based on the publicly available data from the online user community are important. In order to improve the accuracy of existing predicting model, we propose a weighted related networks approach to account for the contributions from related products in the same brand (supporting) and category (competing), plus other factors including the claimed past purchases by the user members, the number of reviewers and the relative prices. Three scenarios of such an approach are implemented, compared and tested with a baseline model that is developed based on a traditional approach in literature. The results show that categorical relationship impacts the online sales more significantly in general.
Keywords: sales prediction, sales network, brand and category, online marketing