You are on page 1of 29

“E-Commerece customer

recommendation system”
Presented by :
K. Venkata Rao(9920004620)

P. Sumanth Kumar Reddy(9920004637)


P. Girish Kumar Reddy(9920004656)
M. Chethan Kumar(9920004660)

UNDER THE GUIDANCE OF


DR.A.SARAVANAN Team Name…..

@Team Beta
OUT LINE

1. INTRODUCTION
2. OVERVIEW
3. RECOMMENDER SYSTEM MODEL
4. RECOMMENDATION ALGORITHMS
5. CHALLENGES OF RECOMMENDER SYSTEMS
6. CHALLENGES OF MOBILE COMMERCE
7. CONCLUSION
8. REFERENCES
INTRODUCTION

The Paradox of Choice

 Overwhelming number of options to consider.


 Lack of effective system support in making decisions
 Too many options can make your visitors too confused and undecided

Only 10%of products on an online retail store garner 75% of page views
OVERVIEW

 What are recommendation systems?


• A recommendation system provides information or items that
are likely to be interest to user in an automated fashion.
• Recommendation systems help match users with items….
WHY DO WE NEED RECOMMENDATION SYSTEMS?

 # Value for the Customer


 - Find things that are interesting
 - Narrow down the set of choice
 - Help explore the space of options
 - Reduce cognitive load on users
 # Value for the provider
 - Additional and unique personalized service for the customer
 - Increase trust and customer loyalty.
 - increase sales, click through rates etc.
 - Opportunities for promotion , persuasion
 - Obtain more knowledge about customers
EVERYDAY EXAMPLES OF RECOMMENDATION SYSTEMS….
MORE
EXAMPLES….

 Netflix predicts other


“Movies You love” based on
past numeric ratings (1-5 stars)
 Recommendations drive 60%
of Netflix’s DVD rentals
MORE EXAMPLES….
MORE
EXAMPLES...
RECOMMENDER SYSTEM
MODEL

Automatically identify item of interest to users (focus of talk)


Candidate Generation

Filters: near duplicates, already seen, dismissed


Filtering

Rank Recommendations based on temporal, geo-location and personalization

User Feedback
RECOMMENDATION
ALGORITHMS

 # Collaborative filtering. (CF)


 Hypothesis : Similar users tend to like similar items Two forms of CF
 - Item-based collaborative filtering
 - User-based collaborative filtering

 # Data Collection Methods


 - Explicit feedback
 Example: ratings, dismiss
 - Implicit feedback
 Example: number of views , purchases
DATA REPRESENTATION

 # Items :i1, i2, i3,…….in


 # u1, u2, u3…….un

 Example of User /Movie Ratings Matrix:

Alice Bob Charlie Dave

Harry Potter…. 3 5 2 3

American Pie 4 4 2 -
Twilight….. 5 1 - -
Matrix - 1 1 5
A NAÏVE RECOMMENDATION SYSTEM

 1. Aggregate ratings for each time


 2. Recommend item with maximum rating

score(i,u) = f(i) n= ∑uэusers ratings(u,i)

 Does everybody like Harry Potter movies?


Historical information about users is important!
ITEM-BASED COLLABORATIVE FILTERING

 # Predict user’s rating for an item I based on his rating for the item
 # Given a user u with I(u) preferred items

Score(i,u) = ∑jэl(u) rating (u,j) . Sim(I, j)

Rating Similarity
provided by between items
user u for item j i and j
EXAMPLE : ITEM-BASED CF

 # Given user with ratings for item X and Y


Harry Potter (X) The Matrix (Y)

rating 0.8 0.3

# Items N and S with similarities

Item Harry Potter (X) The Matrix (Y)


The Chronicles of Narnia 1.0 0.3
(N)
Star Wars (S) 0.2 0.8

Source(u,N) = 1.0*0.8 + 0.3*0.3 = 0.89


Source(u,S) = 0.2*0.8 + 0.3*0.8 = 0.4
COMPUTING SIMILARITY BETWEEN ITEMS

Cosine Similarity
- Items are represented as u-dimensional vectors over user space
- Similarity is cosine of the angle between two vectors
- Score ranges between 1 (perfect) and -1 (opposite)
EXAMPLE: 2
users

Items User 1 User 2

A 0.8 0.45

B 0.4 0.8

C 0.3 0.3
JACCARD SIMILARITY MEASURE

Measures the similarity between finite sample sets

J(A,B) = |AՈB|/|AՍB|

Defined as the size of intersection divided by the size of the union of the sample sets
Sample sets
Sample sets of Items :
A ={Item1,Item3,Item6}
B ={Item1,Item2,Item6}

J(A,B) = 2/4 = 0.5


USER BASED COLLABORATIVE FILTERING

# K – nearest neighbors (KMN)


- Group users into different clusters
# Hard clustering
# Soft clustering

Users Clusters Items


CONTENT-BASED RECOMMENDATION

Collaborative filtering does not require my information about the items


- However, it might be reasonable to exploit such information
- E.g. Recommend fantasy novels to people who liked fantasy novels in the post

What do we need?
- Some information about the available item such as the genre(content)
- Some sort of user profile describing what the user likes
CONTENT REPRESENTATION AND ITEM SIMILARITIES

Title Genre Author Type Price Keywords


The Night of the Gun Memoir David Paperback 29.00 Press and journalism,
Carr dr-ugaddiction,
Personal memories,
Ne-w york

The Laco Reader Fiction, Brunonia Hardcover 49.90 American


Mystery Barry contemporaryfiction,
detective, historical
Into the Fire Romance, Suzanne Hardcover 45.90 American fiction,
Suspense Brock- Murder,
mann Neo-nazism
…… Fiction, Brunonia Paperback 25.65 Detective, murder,
Suspense Barry, New york
Ken
Follet,…
HYBRID RECOMMENDER SYSTEMS

- Combination of collaborative filtering and content based filtering


- Idea of crossing two or more implementations
- Hybrid features
. Social Features
Movies liked by user
. Content features
Dramas liked by user
. Hybrid features
User who like many movies that are dramas
CHALLENGES AND INTERESTING PROBLEMS
RECOMMENDER SYSTEMS

Data sparsity
- Users rarely purchase, rate or click
The more you see the less you know
- Increasing users or items increase the dimensions we need to learn Cold –
start problem
- No historical information for new users or items Scalability
- Increase in the size of matrix
CHALLENGES IN DESIGNING RECOMMENDER
SYSTEMS FOR MOBILE USERS

Size of the display, small screen devices


Limited input and interaction capabilities
Mobile users have shorter browsing sessions
Lack of standardization of the browsing tools
Cost of interaction

Exclusive characteristics :
Location awareness
Ubiquity
Data set

 There are lots of data set available for Recommendation System:


 1. Movielens 100K, 1M , 10M, 20M dataset for movie
 2. Film Trust data set for movie
 3. MiniFIlm Dataset for movie
 4. HyRec2011 dataset for movie
 5. Book-Crossing dataset for book
 6. Jester for joke
 7. Last.fm dataset for movie
 8. Netflix and etc.
 In this we are using Netflix, amazon and movie dataset.
CONCLUSION

# Recommender systems are a huge success in E-commerce sites


# Recommendation systems in mobile commerce have to overcome obstacles
# Mobile devices coupled with Recommender systems would be key tools for business
applications
QUESTION 1

In item based collaborative filtering, based on the user’s previous rating, recommend the most appropriate item to the user
A

User A Strawberries Oranges


Rating 0.8 0.2

Similarity with previously purchased items:


Item Strawberries Oranges
Blueberry(B) 1.0 O.3
Tangerine(T) 0 0.9
Score (u,B) = 0.8*1.0+0.2*0.3 = 0.86
Score (u,T) = 0.8*0+0.2*0.9 = 0.18
The item blueberry is recommended to the user as the score for blueberry is higher
REFERENCES :

 Chengzhi Liu, Caihong Sun and Meigi Fang. The design of an open hybrid
recommendation system for mobile commerce, Communication Technology, 2008.
ICCT 2008. 11th IEEE International Conference on E-ISBN: 978-1-4244-2251-7.
 Azene Zenebe, Ant Ozok and Anthony F. Norcio, Personalized Recommender Systems
in e commerce and m-commerce: A comparitive Study, 11th International Conference
on Human Computer Interaction.
 Badrul Sarwar, George Karypis, Joseph Konstan and John Riedl. Analysis of
recommendation algorithms for e commerce, EC 00 Proceedings of the 2nd ACM
conference on Electronic commerce, ISBN:1 58113-272-7.
 Amund Tveit, Peer to peer based Recommendation for mobile-commerce, ACM
Mobile Commerce Workshop,2001.

You might also like