18. The Features
Direct Match
• Match
Count
• Last Word
Match
• Brand
Matching
Ratios
• Search
Term Hit
Ratio
• Target Hit
Ratio
Length
• Search
Term
• Title
• Description
Disconnected
• Word
Power
Scores
• Synonym
Scores
19. The Document Term Matrix
When all you have is a Home Depot hammer . . .
20. Text Mining Features with doc2vec – cosine distances
1.Distance (Product Title + Product Description vs. Search Term)
2. Categorical Data: Distance (Product Title vs. Product Title)