Huseyin Cenk Ozmutlu

U.S. versus European web searching trends

As the Web is becoming a worldwide phenomenon we need to understand what searching trends are emerging across different global regions. Are there regional differences in Web searching? What are the differences between searching by the United States population compared to Europeans? As part of a body of research studying these questions, we have analyzed two data sets culled from more than one million queries submitted by more than 200,000 users of the Excite Web search engine collected in May 2001 and the FAST Web search engine (All theWeb.com), collected in February 2001.We compare the searching behavior of largely European FAST Web search engine users (mostly German) with Excite Web search engine users who are largely U.S. This comparative study shows differences in Web searching by U.S. and European users. Specifically, the results suggest some differences in the topics searched and searching behaviors.

A day in the life of web searching: an exploratory study

Understanding Web searching behavior is important in developing more successful and cost-efficient Web search engines. We provide results from a comparative time-based Web study of US-based Excite and Norwegian-based Fast Web search logs, exploring variations in user searching related to changes in time of the day. Findings suggest: (1) fluctuations in Web user behavior over the day, (2) user investigations of query results are much longer, and submission of queries and number of users are much higher in the mornings, and (3) some query characteristics, including terms per query and query reformulation, remain steady throughout the day. Implications and further research are discussed.

Multimedia web searching trends: 1997-2001

Multimedia is proliferating on Web sites, as the Web continues to enhance the integration of multimedia and textual information. In this paper we examine trends in multimedia Web searching by Excite users from 1997 to 2001. Results from an analysis of 1,025,910 Excite queries from 2001 are compared to similar Excite datasets from 1997 to 1999. Findings include: (1) queries per multimedia session have decreased since 1997 as a proportion of general queries due to the introduction of multimedia buttons near the query box, (2) multimedia queries identified are longer than non-multimedia queries, and (3) audio queries are more prevalent than image or video queries in identified multimedia queries. Overall, we see multimedia Web searching undergoing major changes as Web content and searching evolves.

Multitasking Web Searching and Implications for Design.

This paper presents findings from a study of users multitasking searches on Web search engines. A users single session with a Web search engine may consist of seeking information on single or multiple topics. Limited research has focused on multitasking search and the implications for Web design. Incidence of multitasking search by AlltheWeb.com and Excite Web search engine users were filtered from transaction logs. Findings include: (1) multitasking Web searches are a noticeable user behavior, one tenth of Excite users and one third of AlltheWeb.com users conducted multitasking searches, (2) multitasking search sessions are longer than regular search sessions in terms of queries per session and duration, (3) both Excite and AlltheWeb.com users search for about three topics per multitasking session and submit about 4–5 queries per topic, and (4) there is a broad variety of search topics in multitasking search sessions. The implications of our findings for Web design and further research are discussed.

Are people asking questions of general Web search engines

Recent studies show that many Web users only submit short queries and conduct short search sessions. This paper examines aspects of users’ attempting longer more complex queries. Web search services such as Ask Jeeves – publicly accessible question and answer (Q&A) search engines – encourage queries in question or request format. In light of this trend, this study examines whether general Web queries are shifting towards a more question/request format. Previous studies show that some users were submitting question or request format queries to general non‐Q&A Web search engines. This paper re‐examines this issue by analysing large‐scale Web query data from two different (US and European) Web query data sets, including 1.2 million Excite queries (www.excite.com) and 1.2 million AlltheWeb.com (http://AlltheWeb.com) queries from 2001.

A study of multitasking Web search

We present findings from a study of users performing multitasking searches on Web search engines. A users single session with a Web search engine may consist of seeking information on single or multiple topics. Limited research has focused on multitasking search query session. The main objective of the study is to provide a detailed analysis of multitasking sessions. FAST search sessions were used to analyze multitasking Web searches. Findings include: (1) almost one third of FAST users perform multitasking Web searching; (2) multitasking sessions often included more than three topics per session; (3) multitasking sessions are longer in duration than regular searching sessions; and (4) most of the topic in multitasking searches were on general information, computers and entertainment.

Zone recovery methodology for probe-subset selection in end-to-end network monitoring

To predict the delay between a source and a destination as well as to identify anomalies in a network, it is possible to monitor the network continuously by sending probes between all sources and destinations. However, it is of prime importance to keep the number of probes to a minimum and yet be able to predict the delays and identify anomalies reasonably. We state and solve a mathematical programming problem, namely the zone recovery methodology (ZRM), to select an optimal subset of ping-like probes to monitor networks where the topology and routing information are not known. A polynomial-time heuristic is developed. The application of ZRM on randomly generated topologies yielded 73.55% reduction in the number of monitored paths on average. In other words, networks can be successfully monitored using only 26.45% of the available probes. Moreover, the performance of ZRM increases (percentage of the monitored paths decreases) as the size of the topology increases.

Character n-gram application for automatic new topic identification

We used the character n-gram method to predict topic changes in search engine queries.We obtained more successful estimations than previous studies, and made remarkable contributions.We compared the character n-gram method with the Levenshtein edit-distance method.We analyzed ASPELL, Google and Bing search engines as pre-processed spelling correction methods.We conclude that Google could be used as a pre-processed spelling correction method. The widespread availability of the Internet and the variety of Internet-based applications have resulted in a significant increase in the amount of web pages. Determining the behaviors of search engine users has become a critical step in enhancing search engine performance. Search engine user behaviors can be determined by content-based or content-ignorant algorithms. Although many content-ignorant studies have been performed to automatically identify new topics, previous results have demonstrated that spelling errors can cause significant errors in topic shift estimates. In this study, we focused on minimizing the number of wrong estimates that were based on spelling errors. We developed a new hybrid algorithm combining character n-gram and neural network methodologies, and compared the experimental results with results from previous studies. For the FAST and Excite datasets, the proposed algorithm improved topic shift estimates by 6.987% and 2.639%, respectively. Moreover, we analyzed the performance of the character n-gram method in different aspects including the comparison with Levenshtein edit-distance method. The experimental results demonstrated that the character n-gram method outperformed to the Levensthein edit distance method in terms of topic identification.

Using conditional probabilities for automatic new topic identification

Purpose – One of the most important dimensions of search engine user information seeking behaviour is content‐based behaviour. One of the main elements in developing a personalised intelligent search engine is new topic identification. The purpose of this study is to perform automatic new topic identification in search engine transaction logs using conditional probabilities of new topic arrivals.Design/methodology/approach – Sample data logs from FAST (currently owned by Yahoo!) and Excite (currently owned by IAC Search & Media) are used in the study. Conditional probabilities of new topic arrivals and topic continuations given query category are used to estimate new topic arrivals.Findings – The findings of this study show that the conditional probability approach reduced overestimation of topic shifts, increasing some performance measures to their highest ever value compared to previous studies. A straightforward procedure such as the conditional probability approach can be as successful as, and for som...

A Monte-Carlo simulation application for automatic new topic identification of search engine transaction logs☆

Abstract One of the most important dimensions of Web user information seeking behavior and search engine research is content-based behavior, and limited research has focused on content-based behavior of search engine users. The purpose of this study is to perform automatic new topic identification in search engine transaction logs using Monte-Carlo simulation. Sample data logs from FAST and Excite are used in the study. Findings show that Monte-Carlo simulation for new topic identification yields satisfactory results in terms of identifying topic continuations; however, the performance measures regarding topic shifts should be improved.