Web retrieval: Techniques for the aggregation and selection of queries and answers

✍ Scribed by David Nettleton; Ricardo Baeza-Yates

• Publisher: John Wiley and Sons
• Year: 2008
• Tongue: English
• Weight: 181 KB
• Volume: 23
• Category: Article
• ISSN: 0884-8173
• DOI: 10.1002/int.20316

✦ Synopsis

In this paper we present a set of techniques for grouping and aggregating queries and search results, in the context of an Internet "search engine." (1) In the case of the initial grouping of the queries, we consider the Fuzzy c-Means 1 and Kohonen SOM 2 techniques. It is proposed that FCM may be more adequate than k-Means 3 for the grouping of certain data types. We evaluate how we can use FCM to calculate the fuzzy membership grades for a set of Web queries and their corresponding results. (2) In the case of the aggregation of data from different information sources (the clustering techniques), we will consider weighted ordered weighted averaging (WOWA). 4 WOWA is used to choose the most adequate cluster and to identify the historical query in that cluster, which is most similar to a new query. We will see that the WOWA operator offers a wide flexibility for data processing.