Topic-sensitive pagerank: A context-sensitive ranking algorithm for web search

Abstract

Abstract-The original PageRank algorithm for improving the ranking of search-query results computes a single vector, using the link structure of the Web, to capture the relative "importance" of Web pages, independent of any particular search query. To yield more accurate search results, we propose computing a set of PageRank vectors, biased using a set of representative topics, to capture more accurately the notion of importance with respect to a particular topic. For ordinary keyword search queries, we compute the topicsensitive PageRank scores for pages satisfying the query using the topic of the query keywords. For searches done in context (e.g., when the search query is performed by highlighting words in a Web page), we compute the topic-sensitive PageRank scores using the topic of the context in which the query appeared. By using linear combinations of these (precomputed) biased PageRank vectors to generate context-specific importance scores for pages at query time, we show that we can generate more accurate rankings than with a single, generic PageRank vector. We describe techniques for efficiently implementing a large-scale search system based on the topic-sensitive PageRank scheme.

PageRank and Random Walks on Graphs

“…In this paper, we will use a more general version of PageRank, called personalized PageRank, introduced by Jeh and Widom [7] (also see Haveliwala [6]). The personalized PageRank pr α (s) depends on two parameters, the jumping constant α and a seed s. A seed can be viewed as a vertex or a probability distribution on vertices.…”

Detecting Sharp Drops in PageRank and a Simplified Local Partitioning Algorithm

“…PageRank vectors whose starting vectors are concentrated on a smaller set of vertices are often called personalized PageRank vectors. These were introduced by Haveliwala [5], and have been used to provide personalized search ranking and context-sensitive search [2,4,6]. We will consider PageRank vectors whose starting vectors are equal to the indicator function 1 v for a single vertex v. The vertex v will be called the starting vertex, and we will use the notation pr(α, v) = pr(α, 1 v ).…”

Classifier Combination Using Random Walks on the Space of Concepts

“…The intuition behind this approach is to model the behaviour of a "random surfer" that with probability (1−α) gets bored and makes a jump to an arbitrary site. An extension to this model -known as personalization-consists on replacing e T /n by v T : a distribution over states reflecting the preferences of each particular user [6].…”