comp423 intelligent agents. recommender systems two approaches – collaborative filtering based on...

10
COMP423 Intelligent Agents

Upload: todd-shelton

Post on 16-Dec-2015

217 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

COMP423Intelligent Agents

Page 2: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

Recommender systems

• Two approaches– Collaborative Filtering

• Based on feedback from other users who have rated a similar set of items in the past

– Content based filtering (e.g SmartMuseum)• Based on how well the contend of the target item matches the

user’s preferred content pattern, which is learnt from the user’s own past ratings and the content pattern of the rated items.

– Hybrid

Page 3: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

User-based Collaborative Filtering

• Nearest Neighbor Collaborative Filtering– Calculate user similarities• Pearson’s correlation

– Define the effective neighborhood– Computer the predicted ratings

𝑟𝑘𝑒𝑛 , 𝑙𝑒𝑒=

∑𝑖=1

𝑛

(𝐾 𝑖−𝜇𝐾𝑒𝑛)(𝐿𝑖−𝜇𝑙𝑒𝑒)

𝑛𝜎𝐾𝑒𝑛𝜎 𝐿𝑒𝑒

+

The correlation of two users ken and lee, they both ranked n items K(1..n) L (1..n)

Prediction on Ken’s ranking for m

Page 4: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

Item-based Collaborative filtering

• Item ranking Matrix• Item vectors: the columns• Item similarity– Pearson’s Correlation– Cosine similarity– Adjusted Cosine similarity

Page 5: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

typical Collaborative Filtering

• Memory based collaborative filtering– Nearest-neighbor based – User similarity– Item similarity

• Clustering for collaborative filtering– Kmeans– HAC– Naïve Bayes clustering

– Group oriented, less personalized, can be addressed by reducing cluster size

Page 6: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

Content based filtering

• Content– Features: • Movie: directors, actor/actress, producers., editors,

distributors, editors, keywords, review, ….• Text recommendation: a set of extracted keywords

Classification problem

Page 7: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

Hybrid• Collaborative filtering:

– Require other users rating data (cold start problem)– Can do cross domain – Non-transitive association problem: users are linked by

common items and items are linked by common users.

• Content Based– Require one user’s rating data– Require item’s content data– Not cross domain

• Sequential Hybridization

• Combinational Hybridization

Page 8: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

Evaluation• Binary: change rates to positive or negative

– Precision– Top N precision– Recall– F-measure– MAP: consider ranking, precision, recall

• Mean of the Average Precision for all queries• Average Precision: the mean of the precision when each relevant

document is retrieved. (M: No of relevant documents)

• Average precision is roughly the area under the precision and recall curve

Page 9: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

Evaluation

• Consider ranking score

• MAE: mean absolute error

Page 10: COMP423 Intelligent Agents. Recommender systems Two approaches – Collaborative Filtering Based on feedback from other users who have rated a similar set

Research projects

• Recommender systems combined with personalized search– Building profile from click through data– Query expansion based on profile

• Two way recommendation– Online dating systems

• Knowledge-based, Personalized recommendation