Professional Documents
Culture Documents
presentation
BITS Pilani
Pilani Campus
N.MEHALA
FACULTY,CS/IS GROUP
INFORMATION RETRIEVAL
CS F469
Second Semester 2014-15
4/14/2015
CS F469
2
BITS Pilani, Deemed to be University under Section 3 of UGC Act, 1956
Search
Items
Products, web sites, blogs, news items,
3
BITS Pilani, Pilani Campus
Example applications
4
BITS Pilani, Pilani Campus
5
BITS Pilani, Pilani Campus
6
BITS Pilani, Pilani Campus
As Prediction
Rating prediction, i.e., predict the rating score
that a user is likely to give to an item that s/he
has not seen or used before. E.g.,
rating on an unseen movie. In this case, the
utility of item s to user u is the rating given to s
by u.
Item prediction, i.e., predict a ranked list of items
that a user is likely to buy or use.
8
BITS Pilani, Pilani Campus
Content-Based
Recommender System
10
BITS Pilani, Pilani Campus
Plan of action
Item profiles
likes
build
recommend
match
Red
Circles
Triangles
User profile
11
BITS Pilani, Pilani Campus
Content-Based Recommendation
In content-based recommendations the system
tries to recommend items that matches the User
Profile.
The Profile is based on items user has liked in the
past or explicit interests that s/he defines.
A content-based recommender system matches
the profile of the item to the user profile to decide
on its relevancy to the user.
12
BITS Pilani, Pilani Campus
Simple Example
update
Read
New books
Match
Recommender
Systems
User Profile
User Profile
recommendation
13
BITS Pilani, Pilani Campus
Advantages
No need for data on other users
No cold-start or sparsity problem
Able to recommend to users with unique tastes
Able to recommend new and unpopular item
No first-rater problem
14
BITS Pilani, Pilani Campus
Collaborative Filtering
Consider user c
Find set D of other users whose ratings are
similar to cs ratings
Estimate users ratings based on ratings of
users in D
Similar
Estimate
Ratings
Ratings
16
BITS Pilani, Pilani Campus
Hybrid Methods
Content-based and collaborative methods have
complementary strengths and weaknesses
Combine methods to obtain the best of both
Various hybrid approaches:
Apply both methods and combine recommendations
Use collaborative data as content
Use content-based predictor as another collaborator
18
BITS Pilani, Pilani Campus
19
BITS Pilani, Pilani Campus
Recommender Systems
Goal recommend services
or items to user
Input - user preferences
defined as a profile
Output - ranked items based
on user preferences
Methods variety of
methods, IR, ML
Query Recommender
Content based or content ignorant or hybrid
recommender for queries using information
about documents
21
BITS Pilani, Pilani Campus
22
BITS Pilani, Pilani Campus
23
BITS Pilani, Pilani Campus
14.04.2015
Recommend a vacation
Winter vs. summer
Recommend a purchase (e-retailer)
Gift vs. for yourself
24
BITS Pilani, Pilani Campus
14.04.2015
25
BITS Pilani, Pilani Campus
26
BITS Pilani, Pilani Campus
Evaluating Recommendations
Precision
Accuracy of predictions
Compare predictions with known ratings
Recommendation Quality
Top-n measures (e.g., Breese score)
Item-Set Coverage
Number of items/users for which system can make
predictions
27
BITS Pilani, Pilani Campus
Evaluating Recommendations
Receiver operating characteristic (ROC)
Tradeoff curve between false positives and false
negatives
Predicted
class
Cat
Actual
class
Dog
Rabbit
Cat
Dog
Rabbit
11
5 true positives
(actual cats that
were correctly
classified as cats)
2 false positives
(dogs that were
incorrectly labeled
as cats)
17 true negatives
3 false negatives (all the remaining
animals, correctly
(cats that were
incorrectly marked classified as noncats)
as dogs)
28
BITS Pilani, Pilani Campus
29
BITS Pilani, Pilani Campus