Recommendation System
@anhmv
Collaborative filtering
Introduction
Recommendation system is an information filtering system that attempt to present the user information they interested in.
1. Collaborative filtering
2. Content-based filtering
3. Hybrid recommendation system
Applications
Amazon
Spotify
Let's get started
Orders History
Vectors
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Dimension
Vector
Vectors
= (1,
1,
1,
1,
1,
1,
1,
= (1,
1,
1,
1)
1,
1,
1,
1,
1)
0,
0,
0,
0,
0,
0,
0)
= (0,
0,
0,
0,
0,
0)
= (0,
0,
0,
0,
0,
0)
= (0,
0,
0,
0,
0,
0,
0,
= (0,
0,
0,
0,
0,
0,
0)
0,
0,
0,
0,
0,
Similarity Calculation
Similarity Matrix
1
2
3
4
5
6
1
2
3
4
5
6
Cosine Similarity - Dot Product
similarity = cos(\theta)= \frac{A \cdot B}{\|A\| \|B\|} = \frac{\displaystyle\sum_{i=1}^{n} A_i B_i}{\displaystyle\sum_{i=1}^{n}\sqrt A^2_i \displaystyle\sum_{i=1}^{n}\sqrt B^2_i}
Similarity Matrix
1
2
3
4
5
6
1
2
3
4
5
6
0
0
0
0
0.707
0
0
0
0
0.707
0
0
0.408
0.288
0
0
0.408
0.288
0
0
0.408
0.408
0.408
0.408
0
0
0.353
0.353
0
0
K-Nearest Neighbor
1
2
3
4
5
6
1
2
3
4
5
6
0
0
0
0
0.707
0
0
0
0
0.707
0
0
0.408
0.288
0
0
0.408
0.288
0
0
0.408
0.408
0.408
0.408
0
0
0.353
0.353
0
0
K-Nearest Neighbor
1
2
4
0.707
0.288
0.353
6
K = 3
Only get 3 most similar users
Remove bought items
1
2
4
0.707
0.288
0.353
6
K = 3
Only get 3 most similar users
0.288
0.288
0.353
Calculate final score
6
0.288
0.288
0.353
Other Similarity Measures
Problems?
1. Cold start
2. Scalability
3. Sparsity
Thank You
Recommendation System
By Anh Mạc Văn