Collaborative multi-level embedding learning from reviews for rating prediction

Keyphrases

• Rating Prediction 100%
• Embedding Learning 100%
• Multi-level Embedding 100%
• Contextual Information 40%
• Local Context 40%
• Word Embedding 40%
• Stochastic Gradient Descent 20%
• Irrelevant Information 20%
• Historical Review 20%
• Item Representation 20%
• Embedding Model 20%
• Gradient Ascent 20%
• Model Factor 20%
• Joint Optimization Problem 20%
• Latent Factor Model 20%
• Review Text 20%
• User Representation 20%
• User Ratings 20%
• Review Representation 20%
• Competitive Method 20%
• Strict Equivalence 20%
• Item Embedding 20%
• Matrix Factorization Model 20%
• Representation Equivalent 20%
• User Embedding 20%

Computer Science

• Context Information 100%
• Word Embedding 100%
• Optimization Problem 50%
• Historical Review 50%
• Matrix Factorization Model 50%