Table 8 Average relative change in F1 score using C+M for property crimes over all test months (Aug-Dec) in each city
From: Enhancing short-term crime prediction with human mobility flows and deep learning architectures
Model | Bal | Min | Aus | Chi |
|---|---|---|---|---|
HALR | 6.1% | 3.6% | 1.6% | 2.6% |
GRU | 5.3% | 4.6% | 1.3% | 3.4% |
Attn | 5.4% | 2.7% | 0.5% | 2.8% |
GCN | −0.6% | −1.2% | 0.4% | 0.2% |
GGConv | 4.0% | 1.4% | 2.4% | 0.6% |
NbConv | 5.8% | 4.7% | 2.1% | 4.7% |