Learning geo-contextual embeddings for commuting flow prediction

Computer Science

• Spatial Correlation 100%
• Contextual Information 100%
• Attention (Machine Learning) 66%
• Experimental Result 33%
• Predictive Power 33%
• Machine Learning 33%
• Learning Framework 33%
• Embedding Space 33%
• Conventional Model 33%
• Demand Characteristic 33%
• Complex Pattern 33%
• Gradient Boosting 33%
• Multitask Learning 33%

Keyphrases

• Effective Representation 20%
• Multi-task Learning Framework 20%
• Graph Attention Network 20%
• Infrastructure Use 20%
• Urban Public Policy 20%