An extension of the gravity model
针对传统引力模型忽略购物设施内停留时间导致市场份额估计偏差的问题，提出通过引入额外距离参数来扩展模型，仅需顾客邮编信息即可实现，并用购物中心数据验证了有效性。
We introduce a practically useful extension of the gravity model. When specifying the distance decay function, the basic gravity model and its variants use actual distance or travel time. But, in reality, travel time to a retail outlet is only a fraction of the time spent on shopping trips. Hence, ignoring “time spent in facilities” may bias market share estimates in a competitive environment. To address this issue and by-passing the need to know time spent on facilities, we propose a simple extension of the gravity model by introducing an extra distance/travel time parameter. For its implementation, zip code (or neighborhood) information of intercepted patrons at facilities is sufficient to determine the extra distance making it practically appealing. To illustrate the estimation and also empirically validate the proposed approach, we analyze a shopping malls application.