Clustering WebSearch Results Using TransductionBased Relevance Model

1
Clustering Web-Search Results Using
Transduction-Based Relevance Model

• Lurong Xiao and Edward Hung
• Hong Kong Polytechnic University
• csehung_at_comp.polyu.edu.hk

2
Outline

• Motivation
• Transduction-based Clustering Algorithm (TCA)
• Result preprocessing
• Similarity measurement
• Transduction-based Relevance Model (TRM)
• Clustering Algorithm
• Performance Evaluation
• Conclusion

3
Motivation

• Existing search engines return a long list of
  ranked results
• Keyword jaguar

4
(No Transcript)
5
(No Transcript)
6
Challenges

• Search engines return titles, snippets and links
• Time consuming to download and process original
  results
• Short texts ? hard to extract reliable features
  and produce good clusters
• Online ? need fast response time
• Clusters should have readable descriptions
• No study on relationship between results and
  distributions

7
Our Approach

• Transduction-based Relevance Model (TRM)
• No assumption on distribution
• Determine relevance (relationship) between two
  results

8
Our Approach

• Clustering algorithm
• Clustering based on relevance
• Compute Importance value of each result (how
  closely other results are relevant to it?)
• An important result ? cluster representative
• Importance of results relevant to them are
  suppressed
• Repeat to pick important results above threshold
• Assign other results to most relevant clusters
• Results with very low relevance ? outliers (a
  special cluster with miscellaneous topics)
• Adjust importance and relevance thresholds ?
  determine cluster number and granularity of
  clustering

9
Transduction-based Clustering Algorithm

• Result preprocessing
• Similarity measurement
• Transduction-based relevance model
• Clustering algorithm

10
1. Result preprocessing

• User query ? a list of ranked results
• First n results X x1, x2, , xn
• Term set of m terms
• an m x 1 cell array all terms in titles and
  snippets (except non-term tokens and frequent
  terms)
• m x n histogram matrix H(hi,j)mxn where H aHt
  (1- a)Hs
• H(k,i) weighted occurrence of k-th term in xi

11
2. Similarity measurement

• Normalized tf-idf weighted vectors
• Term frequency
• Inverse document frequency
• Weight of k-th term in i-th result
• cosine similarity
• Distance between results xi, xj
• n x n distance matrix D

12
3. Transduction-based Relevance Model (TRM)

• Construct a KNN graph based on distance matrix D
• Result xi ? node
• xj is a KNN of xi ? xj in N(xi) ? edge (xi, xj)
• Affinity weight matrix W(wi,j)nxn

13
3. Transduction-based Relevance Model (TRM)

• Propagation coefficient matrix Q(qi,j)nxn
• Propagate relevance to get the Relevance matrix
  R(ri,j)nxn
• R Q x R
• ri,i1
• R ? I
• xi has a high relevance to xj if xi is near to
  nodes (e.g. xk) with high relevances to xj

14
3. Transduction-based Relevance Model (TRM)

• Iteratively propagation algorithm
• Iteratively apply until R becomes stable

15
3. Transduction-based Relevance Model (TRM)

• Matrix solution
• Applying matrix equations instead of iteratively
  propagations of relevance values

16
3. Transduction-based Relevance Model (TRM)

• Importance of node xi
• A node is more important iff it has more other
  nodes highly relevant to it

17
4. Clustering Algorithm

• Iteratively find Nc or fewer cluster
  representatives with Imi gt threshold
• Pick node xi with highest importance value Imi
• Avoid to pick nodes highly relevant to xi
• Suppression attenuation
• Assign other nodes to the most relevant cluster
  representatives
• If relevance lt threshold ? outlier or special
  cluster for miscellaneous topics

18
Performance Evaluation

• Query log of June 10, 2007 from google
• Keywords with multiple subtopics and commonly
  used in literature
• jaguar, iraq, java
• a0.5, s1, w00.001, 20-NN graph
• Search results 50, 75, 100
• Relevance threshold 0.001, 0.01
• Importance threshold 2.5, 3, 3.5
• 98.9 accuracy (higher than k-medoids clustering
  88.3)
• Running time 0.3s 0.7s

19
Performance Evaluation

• Higher relevance threshold ? higher accuracy
• More relevant results are usually more likely to
  be correct
• Lower importance threshold ? more clusters ?
  higher accuracy
• Members of new clusters were usually outliers or
  from clusters with low ranks (relevances)
• More data ? more clusters
• Some results were once too small in quantity and
  relevance to form a cluster
• They now become important enough to form their
  own cluster

20
Example on member movement in different cluster
numbers
21
Conclusion

• Transduction-based Clustering Algorithm (TCA)
• Analyze inter-result relationship by propagating
  relevance through local interactions
• Given importance threshold, TCA decides the
  number of clusters automatically
• Search results clustering and outlier detection
• Fast running time and high accuracy, suitable for
  online users