TrajectoryNet: An Embedded GPS Trajectory Representation for Point-based Classification Using Recurrent Neural Networks

Keyphrases

• Global Positioning System 100%
• Point-based 100%
• Recurrent Neural Network 100%
• System Trajectory 100%
• Trajectory Representation 100%
• Feature Space 66%
• Transportation Mode 66%
• System Trace 66%
• Human Activities 33%
• Predictive Power 33%
• Classification Accuracy 33%
• Location-based 33%
• Neural Network Architecture 33%
• Sensory Data 33%
• Heterogeneous Feature Spaces 33%
• Novel Representation 33%
• Basis Expansion 33%
• Segment-based 33%
• Latent Factors 33%
• Low-dimensional Feature Space 33%
• Human Transportation 33%
• Global Positioning System Data 33%
• Maxout 33%
• Urban Computing 33%
• Data-centric 33%
• Trajectory Classification 33%

Computer Science

• System Trajectory 100%
• Recurrent Neural Network 100%
• Global Positioning System 100%
• Feature Space 75%
• Transportation Mode 50%
• Classification Accuracy 25%
• Dimensional Feature 25%
• Neural Network Architecture 25%
• Predictive Power 25%