Figure 2
From: Enhancing short-term crime prediction with human mobility flows and deep learning architectures
Arrange the nearest neighbors set for the target census tract \(s_{1}\) and construct the 2D feature map for historical crimes. In the neighboring set of \(s_{1}\), \(s_{2}\) and \(s_{3}\) is the closest to \(s_{1}\); \(s_{4}\) and \(s_{5}\) are the next closest to \(s_{2}\) and \(s_{3}\) respectively; \(s_{6}\) and \(s_{7}\) are the next closest to \(s_{4}\) and \(s_{5}\); \(s_{8}\) and \(s_{9}\) are the next closest to \(s_{6}\) and \(s_{7}\). Similar process is applied to each of the ten mobility features