how to build a recommender system
DESCRIPTION
This presentation show the method to build a Recommender System with Collaborative FIltering method.TRANSCRIPT
Recommender System
How to build a
Võ Duy TuấnTechnical Director @ dienmay.com
PHP 5 Zend Certified Engineer
Mobile App Developer
Web Developer & Designer
Interest: o PHP
o Large System & Data Mining
o Web Performance Optimization
o Mobile Development
Introduction
Collaborative Filtering
Question & Answer
AGENDA
1. Introduction
APPLICATIONS
• Personalized recommendation• Social recommendation• Item recommendation• Combination of 3 approaches above
AMAZON.COM | BOOKS
PLAY.GOOGLE.COM | APPS
SKILLSHARE.COM | CLASSES
PROCESS DIAGRAM
Preprocessing Data Analysis Adjustment
INPUT OUTPUT
TYPE OF RECOMMENDER SYSTEM
• Collaborative filtering• Content-based filtering• Hybrid
2. Collaborative Filtering
USER & ITEM
ORDER DATA
ORDER DATA (cont.)
ORDER DATA (cont.)
VECTOR & DIMENSION
VECTOR & DIMENSION
VECTORS
VECTORS
SIMILARITY CALCULATION
USER SIMILARITY MATRIX
SIMILARITY CALCULATION
SIMILARITY CALCULATION
SIMILARITY CALCULATION EXAMPLE
K-NEAREST-NEIGHBOR
K-NEAREST-NEIGHBOR
NEIGHBORS’ ORDER
REMOVE BOUGHT ITEMS
CALCULATING FINAL SCORE
OTHER SIMILARITY MEASURES
More at: http://favi.com.vn/wp-content/uploads/2012/05/pg049_Similarity_Measures_for_Text_Document_Clustering.pdf
Problem ?!
COLLABORATIVE FILTERING PROBLEM
• Fail with cold start problemo New Usero New Item
• Performanceo Large Data seto Pre-calculate
PERFORMANCE EXAMPLE
• We have 1,000,000 users (customers)• We sell 10,000 items
- Total of similarity calculating = 1,000,000 x 1,000,000 = 1,000,000,000,000- Each similarity calculate need 0.006s (on my MacBook Pro 2.2GHz Core i7, 8G Ram)
=> We need 1,000,000,000,000 x 0.006 = 6,000,000,000(s)≈ 70,000 days ≈ 191 years
- If store each similarity in 8 bytes, we need = 8,000,000,000,000 bytes≈ 8,000 GB (on Memory or File)
ITEM-TO-ITEM COLLABORATIVE FILTERING (AMAZON.COM )
Download Paper: http://www.cs.umd.edu/~samir/498/Amazon-Recommendations.pdf
ADJUSTMENTS
• Hybrid Recommender System• Sale forecast system• Context of User• Type of Item, Action• External (3rd-party) information.
BOOKS
Programming Collective IntelligenceToby Segaran
Recommender Systems HandbookMany Authors
Big Data For DummiesMarcia Kaufman, Fern Halper
OPEN SOURCES