We establish a methodology that fuses point data with trajectory data leading to trip chains that reflect whether a person has visited key locations. In our study the point data are Wi-Fi data set from sensors at key tourist locations and the trajectories are a small sample of GPS footprints. The approach proves valuable in the context of tour estimation, placing a specific emphasis on determining whether major attractions have been visited. The data fusion approach evaluates the similarity of a sequence of Wi-Fi records with the GPS records which function as a “look-up database”. Based on this, likely missing observations in the Wi-Fi data sequence are estimated. With a Kyoto case study, we demonstrate that the fused data create reasonable tours. Through cross-validation it is verified that missing observations are fairly close to the true location. We then use the fused data with a Recursive Logit approach to estimate parameters that indicate the location attractiveness of touristic areas. The results demonstrate that we can estimate the attractiveness of restaurants and shops to predict tourist routes.

中文翻译：

---

基于 Wi-Fi 传感数据和 GPS 轨迹与京都案例研究的融合，推断旅游模式和旅游区的吸引力

我们建立了一种方法，将点数据与轨迹数据融合，形成反映一个人是否访问过关键地点的行程链。在我们的研究中，点数据是来自主要旅游地点传感器的 Wi-Fi 数据集，轨迹是 GPS 足迹的小样本。该方法在旅游估算方面被证明是有价值的，它特别强调确定主要景点是否已被参观。数据融合方法评估一系列 Wi-Fi 记录与充当“查找数据库”的 GPS 记录的相似性。基于此，估计 Wi-Fi 数据序列中可能缺失的观测值。通过京都案例研究，我们证明融合数据可以创建合理的旅行。通过交叉验证，可以验证缺失的观测值与真实位置相当接近。然后，我们使用融合数据和递归 Logit 方法来估计指示旅游区位置吸引力的参数。结果表明，我们可以估计餐馆和商店的吸引力来预测旅游路线。