# advertising - CS 345 Data Mining Online algorithms Search...

This preview shows pages 1–12. Sign up to view the full content.

CS 345 Data Mining Online algorithms Search advertising

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Online algorithms Classic model of algorithms You get to see the entire input, then compute  some function of it In this context, “offline algorithm” Online algorithm You get to see the input one piece at a time,  and need to make irrevocable decisions along  the way How is this different from the data stream  model?
Example: Bipartite matching 1 2 3 4 a b c d Girls Boys

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Example: Bipartite matching 1 2 3 4 a b c d M = {(1,a),(2,b),(3,d)} is a matching Cardinality of matching = |M| = 3 Girls Boys
Example: Bipartite matching 1 2 3 4 a b c d Girls Boys M = {(1,c),(2,b),(3,d),(4,a)} is a  perfect matching

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Matching Algorithm Problem: Find a maximum-cardinality  matching A perfect one if it exists There is a polynomial-time offline algorithm  (Hopcroft and Karp 1973) But what if we don’t have the entire graph  upfront?
Online problem Initially, we are given the set Boys In each round, one girl’s choices are  revealed At that time, we have to decide to either: Pair the girl with a boy Don’t pair the girl with any boy Example of application: assigning tasks to  servers

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Online problem 1 2 3 4 a b c d (1,a) (2,b) (3,d)
Greedy algorithm Pair the new girl with any eligible boy If there is none, don’t pair girl How good is the algorithm?

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Competitive Ratio For input I, suppose greedy produces  matching M greedy  while an optimal matching is  M opt Competitive ratio =  min all possible inputs I  (|M greedy |/|M opt |)
Analyzing the greedy algorithm Consider the set G of girls matched in M opt  but not in  M greedy

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}

### Page1 / 34

advertising - CS 345 Data Mining Online algorithms Search...

This preview shows document pages 1 - 12. Sign up to view the full document.

View Full Document
Ask a homework question - tutors are online