ValuePick: Towards a Value-Oriented Dual-Goal Recommender System - PowerPoint PPT Presentation

Loading...

PPT – ValuePick: Towards a Value-Oriented Dual-Goal Recommender System PowerPoint presentation | free to download - id: 77d8e6-OGVkM



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

ValuePick: Towards a Value-Oriented Dual-Goal Recommender System

Description:

ValuePick: Towards a Value-Oriented Dual-Goal Recommender System Leman Akoglu Christos Faloutsos OEDM in conjunction with ICDM 2010 Sydney, Australia – PowerPoint PPT presentation

Number of Views:41
Avg rating:3.0/5.0
Slides: 19
Provided by: Polo153
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: ValuePick: Towards a Value-Oriented Dual-Goal Recommender System


1
ValuePick Towards a Value-Oriented Dual-Goal
Recommender System
  • Leman Akoglu Christos Faloutsos

OEDM in conjunction with ICDM 2010
Sydney, Australia
2
Recommender Systems
Traditional recommender systems try to achieve
high user satisfaction
3
Dual-goal Recommender Systems
-value
Trade-off user satisfaction vs. vendor profit
Dual-goal recommender systems try to achieve (1)
high user satisfaction as well as (2)
high-value vendor gain
4
Dual-goal Recommender Systems
vertices ranked by proximity
v253
v162
v261
v327
. . .
query vertex
network-value
5
Dual-goal Recommender Systems
vertices ranked by proximity
v253
v162
v261
v327
. . .
network-value
6
Dual-goal Recommender Systems
network-value
vertices ranked by proximity
v253
v162
v261
v327
Trade-off user satisfaction vs. network
connectivity
. . .
network-value
7
Dual-goal Recommender Systems
  • Main concerns
  • We cannot make the highest value recommendations
  • Recommendations should still reflect users likes
    relatively well

User
8
ValuePick Main idea
  • Carefully perturb (change the order of) the
    proximity-ranked list of recommendations
  • Controlled by a tolerance for each user

?
?
9
ValuePick Optimization Framework
DETAILS
proximity
value
Total expected gain (assuming proximity
acceptance prob.)
tolerance ? 0,1
average proximity score of original top-k
10
ValuePick 0-1 Knapsack
DETAILS
value
We use CPLEX to solve our integer programming
optimization problem
maximum weight W allowed
weight of item i
11
Pros and Cons of ValuePick
  • Cons In marketing, it is hard to predict the
    effect of an intervention in the marketing
    scheme, i.e., not clear how users will respond to
    adjustments
  • Pros
  • Tolerance ? can flexibly (and even dynamically)
    control the level-of-adjustment
  • Users rate same item differently at different
    times, i.e., users have natural variability in
    their decisions.

12
Experimental Setup I
  • Two real networks
  • Netscience collaboration network
  • DBLP co-authorship network
  • Four recommendation schemes
  • No Gain Optimization (? 0)
  • ValuePick (? 0.01, ? 0.02)
  • Max Gain Optimization (? 1)
  • Random
  • value is centrality

13
Experimental Setup II
Simulation steps
  • Given a recommendation scheme s
  • At each step T
  • For each node i
  • Make a set K of recommendations to node i using s
  • Node i links to node j?K with prob.
    proximity(i,j)
  • Re-compute proximity and centrality scores

We use k5 and T30
14
Comparison of schemes
EXPERIMENTS
ValuePick provides a balance between user
satisfaction (high E), and vendor gain (small
diameter).
15
Recommend by heuristic
EXPERIMENTS
Simple perturbation heuristics do not balance
user satisfaction and vendor gain properly.
16
Computational complexity
EXPERIMENTS
Making k ValuePick recommendations to a given
node involves 1 - finding PPR scores
O(edges) 2 - solving ValuePick optimization
w/ CPLEX 1/10 sec. to solve among top 1K nodes
17
Conclusions
  • Problem formulation incorporate the value of
    recommendations into the system
  • Design of ValuePick
  • parsimonious ? single parameter ?
  • flexible ? adjust ? for each user dynamically
  • general ? use any value metric
  • Performance study
  • experiments to show proper trade of user
    acceptance in exchange for higher gain
  • CPLEX with fast solutions

18
  • THANK YOU
  • www.cs.cmu.edu/lakoglu
  • lakoglu_at_cs.cmu.edu

?
About PowerShow.com