How are we searching the World Wide Web? A comparison of nine search engine transaction logs

The Web and especially major Web search engines are essential tools in the quest to locate online information for many people. This paper reports results from research that examines characteristics and changes in Web searching from nine studies of five Web search engines based in the US and Europe. We compare interactions occurring between users and Web search engines from the perspectives of session length, query length, query complexity, and content viewed among the Web search engines. The results of our research shows (1) users are viewing fewer result pages, (2) searchers on US-based Web search engines use more query operators than searchers on European-based search engines, (3) there are statistically significant differences in the use of Boolean operators and result pages viewed, and (4) one cannot necessary apply results from studies of one particular Web search engine to another Web search engine. The wide spread use of Web search engines, employment of simple queries, and decreased viewing of result pages may have resulted from algorithmic enhancements by Web search engine companies. We discuss the implications of the findings for the development of Web search engines and design of online content.     

How doctors search: A study of query behaviour and the impact on search results

Professional, workplace searching is different from general searching, because it is typically limited to specific facets and targeted to a single answer. We have developed the semantic component (SC) model, which is a search feature that allows searchers to structure and specify the search to context-specific aspects of the main topic of the documents. We have tested the model in an interactive searching study with family doctors with the purpose to explore doctors’ querying behaviour, how they applied the means for specifying a search, and how these features contributed to the search outcome. In general, the doctors were capable of exploiting system features and search tactics during the searching. Most searchers produced well-structured queries that contained appropriate search facets. When searches failed it was not due to query structure or query length. Failures were mostly caused by the well-known vocabulary problem. The problem was exacerbated by using certain filters as Boolean filters. The best working queries were structured into 2–3 main facets out of 3–5 possible search facets, and expressed with terms reflecting the focal view of the search task. The findings at the same time support and extend previous results about query structure and exhaustivity showing the importance of selecting central search facets and express them from the perspective of search task. The SC model was applied in the highest performing queries except one. The findings suggest that the model might be a helpful feature to structure queries into central, appropriate facets, and in returning highly relevant documents.     

An analysis of Web searching by European AlltheWeb.com users

The Web has become a worldwide source of information and a mainstream business tool. It is changing the way people conduct the daily business of their lives. As these changes are occurring, we need to understand what Web searching trends are emerging within the various global regions. What are the regional differences and trends in Web searching, if any? What is the effectiveness of Web search engines as providers of information? As part of a body of research studying these questions, we have analyzed two data sets collected from queries by mainly European users submitted to AlltheWeb.com on 6 February 2001 and 28 May 2002. AlltheWeb.com is a major and highly rated European search engine. Each data set contains approximately a million queries submitted by over 200,000 users and spans a 24-h period. This longitudinal benchmark study shows that European Web searching is evolving in certain directions. There was some decline in query length, with extremely simple queries. European search topics are broadening, with a notable percentage decline in sexual and pornographic searching. The majority of Web searchers view fewer than five Web documents, spending only seconds on a Web document. Approximately 50% of the Web documents viewed by these European users were topically relevant. We discuss the implications for Web information systems and information content providers.     

CIDER: Concept-based image diversification,exploration, and retrieval

Many of the approaches to image retrieval on the Web have their basis in text retrieval. However, when searchers are asked to describe their image needs, the resulting query is often short and potentially ambiguous. The solution we propose is to perform automatic query expansion using Wikipedia as the source knowledge base, resulting in a diversification of the search results. The outcome is a broad range of images that represent the various possible interpretations of the query. In order to assist the searcher in finding images that match their specific intentions for the query, we have developed an image organization method that uses both the conceptual information associated with each image, and the visual features extracted from the images. This, coupled with a hierarchical organization of the concepts, provides an interactive interface that takes advantage of the searchers’ abilities to recognize relevant concepts, filter and focus the search results based on these concepts, and visually identify relevant images while navigating within the image space. In this paper, we outline the key features of our image retrieval system (CIDER), and present the results of a preliminary user evaluation. The results of this study illustrate the potential benefits that CIDER can provide for searchers conducting image retrieval tasks.     

Examining the effectiveness of real-time query expansion

Interactive query expansion (IQE) (c.f. [Efthimiadis, E. N. (1996). Query expansion. Annual Review of Information Systems and Technology, 31, 121–187]) is a potentially useful technique to help searchers formulate improved query statements, and ultimately retrieve better search results. However, IQE is seldom used in operational settings. Two possible explanations for this are that IQE is generally not integrated into searchers’ established information-seeking behaviors (e.g., examining lists of documents), and it may not be offered at a time in the search when it is needed most (i.e., during the initial query formulation). These challenges can be addressed by coupling IQE more closely with familiar search activities, rather than as a separate functionality that searchers must learn. In this article we introduce and evaluate a variant of IQE known as Real-Time Query Expansion (RTQE). As a searcher enters their query in a text box at the interface, RTQE provides a list of suggested additional query terms, in effect offering query expansion options while the query is formulated. To investigate how the technique is used – and when it may be useful – we conducted a user study comparing three search interfaces: a baseline interface with no query expansion support; an interface that provides expansion options during query entry, and a third interface that provides options after queries have been submitted to a search system. The results show that offering RTQE leads to better quality initial queries, more engagement in the search, and an increase in the uptake of query expansion. However, the results also imply that care must be taken when implementing RTQE interactively. Our findings have broad implications for how IQE should be offered, and form part of our research on the development of techniques to support the increased use of query expansion.     

Image searching on the Excite Web search engine

A growing body of research is beginning to explore the information-seeking behavior of Web users. The vast majority of these studies have concentrated on the area of textual information retrieval (IR). Little research has examined how people search for non-textual information on the Internet, and few large-scale studies has investigated visual information-seeking behavior with general-purpose Web search engines. This study examined visual information needs as expressed in users’ Web image queries. The data set examined consisted of 1,025,908 sequential queries from 211,058 users of Excite, a major Internet search service. Twenty-eight terms were used to identify queries for both still and moving images, resulting in a subset of 33,149 image queries by 9855 users. We provide data on: (1) image queries – the number of queries and the number of search terms per user, (2) image search sessions – the number of queries per user, modifications made to subsequent queries in a session, and (3) image terms – their rank/frequency distribution and the most highly used search terms. On average, there were 3.36 image queries per user containing an average of 3.74 terms per query. Image queries contained a large number of unique terms. The most frequently occurring image related terms appeared less than 10% of the time, with most terms occurring only once. We contrast this to earlier work by P.G.B. Enser, Journal of Documentation 51 (2) (1995) 126–170, who examined written queries for pictorial information in a non-digital environment. Implications for the development of models for visual information retrieval, and for the design of Web search engines are discussed.     

Querying XML documents using Prolog engines: When is this a good idea?

XML has become a universal standard for information exchange over the Web due to features such as simple syntax and extensibility. Processing queries over these documents has been the focus of several research groups. In fact, there is broad literature in efficient XML query processing which explore indexes, fragmentation techniques, etc. However, for answering complex queries, existing approaches mainly analyze information that is explicitly defined in the XML document. A few work investigate the use of Prolog to increase the query possibilities, allowing inference over the data content. This can cause a significant increase in the query possibilities and expressive power, allowing access to non-obvious information. However, this requires translating the XML documents into Prolog facts. But for regular queries (which do not require inference), is this a good alternative? What kind of queries could benefit from the Prolog translation? Can we always use Prolog engines to execute XML queries in an efficient way? There are many questions involved in adopting an alternative approach to run XML queries. In this work, we investigate this matter by translating XML queries into Prolog queries and comparing the query processing times using Prolog and native XML engines. Our work contributes by providing a set of heuristics that helps users to decide when to use Prolog engines to process a given XML query. In summary, our results show that queries that search elements by a key value or by its position (simple search) are more efficient when run in Prolog than in native XML engines. Also, queries over large datasets, or that searches for substrings perform better when run by native XML engines.     

Using genetic algorithms to evolve a population of topical queries

Systems for searching the Web based on thematic contexts can be built on top of a conventional search engine and benefit from the huge amount of content as well as from the functionality available through the search engine interface. The quality of the material collected by such systems is highly dependant on the vocabulary used to generate the search queries. In this scenario, selecting good query terms can be seen as an optimization problem where the objective function to be optimized is based on the effectiveness of a query to retrieve relevant material. Some characteristics of this optimization problem are: (1) the high-dimensionality of the search space, where candidate solutions are queries and each term corresponds to a different dimension, (2) the existence of acceptable suboptimal solutions, (3) the possibility of finding multiple solutions, and in many cases (4) the quest for novelty. This article describes optimization techniques based on Genetic Algorithms to evolve “good query terms” in the context of a given topic. The proposed techniques place emphasis on searching for novel material that is related to the search context. We discuss the use of a mutation pool to allow the generation of queries with new terms, study the effect of different mutation rates on the exploration of query-space, and discuss the use of a especially developed fitness function that favors the construction of queries containing novel but related terms.     

Characteristics of question format web queries: an exploratory study

Web queries in question format are becoming a common element of a user's interaction with Web search engines. Web search services such as Ask Jeeves – a publicly accessible question and answer (Q&A) search engine – request users to enter question format queries. This paper provides results from a study examining queries in question format submitted to two different Web search engines – Ask Jeeves that explicitly encourages queries in question format and the Excite search service that does not explicitly encourage queries in question format. We identify the characteristics of queries in question format in two different data sets: (1) 30,000 Ask Jeeves queries and 15,575 Excite queries, including the nature, length, and structure of queries in question format. Findings include: (1) 50% of Ask Jeeves queries and less than 1% of Excite were in question format, (2) most users entered only one query in question format with little query reformulation, (3) limited range of formats for queries in question format – mainly “where”, “what”, or “how” questions, (4) most common question query format was “Where can I find………” for general information on a topic, and (5) non-question queries may be in request format. Overall, four types of user Web queries were identified: keyword, Boolean, question, and request. These findings provide an initial mapping of the structure and content of queries in question and request format. Implications for Web search services are discussed.     

Learning from homologous queries and semantically related terms for query auto completion

Query auto completion (QAC) models recommend possible queries to web search users when they start typing a query prefix. Most of today’s QAC models rank candidate queries by popularity (i.e., frequency), and in doing so they tend to follow a strict query matching policy when counting the queries. That is, they ignore the contributions from so-called homologous queries, queries with the same terms but ordered differently or queries that expand the original query. Importantly, homologous queries often express a remarkably similar search intent. Moreover, today’s QAC approaches often ignore semantically related terms. We argue that users are prone to combine semantically related terms when generating queries.We propose a learning to rank-based QAC approach, where, for the first time, features derived from homologous queries and semantically related terms are introduced. In particular, we consider: (i) the observed and predicted popularity of homologous queries for a query candidate; and (ii) the semantic relatedness of pairs of terms inside a query and pairs of queries inside a session. We quantify the improvement of the proposed new features using two large-scale real-world query logs and show that the mean reciprocal rank and the success rate can be improved by up to 9% over state-of-the-art QAC models.     

Distributed search based on self-indexed compressed text

Query response times within a fraction of a second in Web search engines are feasible due to the use of indexing and caching techniques, which are devised for large text collections partitioned and replicated into a set of distributed-memory processors. This paper proposes an alternative query processing method for this setting, which is based on a combination of self-indexed compressed text and posting lists caching. We show that a text self-index (i.e., an index that compresses the text and is able to extract arbitrary parts of it) can be competitive with an inverted index if we consider the whole query process, which includes index decompression, ranking and snippet extraction time. The advantage is that within the space of the compressed document collection, one can carry out the posting lists generation, document ranking and snippet extraction. This significantly reduces the total number of processors involved in the solution of queries. Alternatively, for the same amount of hardware, the performance of the proposed strategy is better than that of the classical approach based on treating inverted indexes and corresponding documents as two separate entities in terms of processors and memory space.     

Using Web structure and summarisation techniques for Web content mining

The dynamic nature and size of the Internet can result in difficulty finding relevant information. Most users typically express their information need via short queries to search engines and they often have to physically sift through the search results based on relevance ranking set by the search engines, making the process of relevance judgement time-consuming. In this paper, we describe a novel representation technique which makes use of the Web structure together with summarisation techniques to better represent knowledge in actual Web Documents. We named the proposed technique as Semantic Virtual Document (SVD). We will discuss how the proposed SVD can be used together with a suitable clustering algorithm to achieve an automatic content-based categorization of similar Web Documents. The auto-categorization facility as well as a “Tree-like” Graphical User Interface (GUI) for post-retrieval document browsing enhances the relevance judgement process for Internet users. Furthermore, we will introduce how our cluster-biased automatic query expansion technique can be used to overcome the ambiguity of short queries typically given by users. We will outline our experimental design to evaluate the effectiveness of the proposed SVD for representation and present a prototype called iSEARCH (Intelligent SEarch And Review of Cluster Hierarchy) for Web content mining. Our results confirm, quantify and extend previous research using Web structure and summarisation techniques, introducing novel techniques for knowledge representation to enhance Web content mining.     

Time series analysis of a Web search engine transaction log

In this paper, we use time series analysis to evaluate predictive scenarios using search engine transactional logs. Our goal is to develop models for the analysis of searchers’ behaviors over time and investigate if time series analysis is a valid method for predicting relationships between searcher actions. Time series analysis is a method often used to understand the underlying characteristics of temporal data in order to make forecasts. In this study, we used a Web search engine transactional log and time series analysis to investigate users’ actions. We conducted our analysis in two phases. In the initial phase, we employed a basic analysis and found that 10% of searchers clicked on sponsored links. However, from 22:00 to 24:00, searchers almost exclusively clicked on the organic links, with almost no clicks on sponsored links. In the second and more extensive phase, we used a one-step prediction time series analysis method along with a transfer function method. The period rarely affects navigational and transactional queries, while rates for transactional queries vary during different periods. Our results show that the average length of a searcher session is approximately 2.9 interactions and that this average is consistent across time periods. Most importantly, our findings shows that searchers who submit the shortest queries (i.e., in number of terms) click on highest ranked results. We discuss implications, including predictive value, and future research.     

Many are Better than One: Algorithm Selection for Faster Top-K Retrieval

Large-scale search engines have become a fundamental tool to efficiently access information on the Web. Typically, users expect answers in sub-second time frames, which demands highly efficient algorithms to traverse the data structures to return the top-k results. Despite different top-k algorithms that avoid processing all postings for all query terms, finding one algorithm that performs the fastest on any query is not always possible. The fastest average algorithm does not necessarily perform the best on all queries when evaluated on a per-query basis. To overcome this challenge, we propose to combine different state-of-the-art disjunctive top-k query processing algorithms to minimize the execution time by selecting the most promising one for each query. We model the selection step as a classification problem in a machine-learning setup. We conduct extensive experimentation and compare the results against state-of-the-art baselines using standard document collections and query sets. On ClueWeb12, our proposal shows a speed-up of up to 1.20x for non-blocked index organizations and 1.19x for block-based ones. Moreover, tail latencies are reduced showing proportional improvements on average, but a resulting dramatic decrease in latency variance. Given these findings, the proposed approach can be easily applied to existing search infrastructures to speed up query processing and reduce resource consumption, positively impacting providers’ operative costs.     

A theoretical framework for defining similarity measures for boolean search request formulations, including some experimental results

Clusters of queries submitted to a given information retrieval system can be used as a basis for an effective method of clustering documents. This indirect procedure of document clustering requires the availability of a similarity measure for queries. Research carried out along these lines has resulted in the development of some methodologies for estimating such query similarities, applicable both in the case of queries characterized by sets of weighted or unweighted index terms and in the case of queries represented by Boolean combinations of index terms. This paper reports the results of further research by the author into a methodology of the latter type, i.e. a methodology for determining the similarity between queries characterized by Boolean search request formulations. The novelty of the presented approach, as compared with the methodology introduced in an earlier paper by the author, is that some relations among index terms are now taken into account. A number of similarity measures for Boolean combinations of index terms are discussed here in some detail. The rationale behind these measures is outlined, and the conditions to be met for ensuring their equivalence are identified. Moreover, the results of an experiment concerning two of the similarity measures introduced are given.     

Real time search on the web: Queries,topics, and economic value

Real time search is an increasingly important area of information seeking on the Web. In this research, we analyze 1,005,296 user interactions with a real time search engine over a 190 day period. Using query log analysis, we investigate searching behavior, categorize search topics, and measure the economic value of this real time search stream. We examine aggregate usage of the search engine, including number of users, queries, and terms. We then classify queries into subject categories using the Google Directory topical hierarchy. We next estimate the economic value of the real time search traffic using the Google AdWords keyword advertising platform. Results shows that 30% of the queries were unique (used only once in the entire dataset), which is low compared to traditional Web searching. Also, 60% of the search traffic comes from the search engine’s application program interface, indicating that real time search is heavily leveraged by other applications. There are many repeated queries over time via these application program interfaces, perhaps indicating both long term interest in a topic and the polling nature of real time queries. Concerning search topics, the most used terms dealt with technology, entertainment, and politics, reflecting both the temporal nature of the queries and, perhaps, an early adopter user-based. However, 36% of the queries indicate some geographical affinity, pointing to a location-based aspect to real time search. In terms of economic value, we calculate this real time search stream to be worth approximately US $33,000,000 (US $33 M) on the online advertising market at the time of the study. We discuss the implications for search engines and content providers as real time content increasingly enters the main stream as an information source.     

SSTS: A syntactic tool for pattern search on time series

Nowadays, data scientists are capable of manipulating and extracting complex information from time series data, given the current diversity of tools at their disposal. However, the plethora of tools that target data exploration and pattern search may require an extensive amount of time to develop methods that correspond to the data scientist's reasoning, in order to solve their queries. The development of new methods, tightly related with the reasoning and visual analysis of time series data, is of great relevance to improving complexity and productivity of pattern and query search tasks. In this work, we propose a novel tool, capable of exploring time series data for pattern and query search tasks in a set of 3 symbolic steps: Pre-Processing, Symbolic Connotation and Search. The framework is called SSTS (Symbolic Search in Time Series) and uses regular expression queries to search the desired patterns in a symbolic representation of the signal. By adopting a set of symbolic methods, this approach has the purpose of increasing the expressiveness in solving standard pattern and query tasks, enabling the creation of queries more closely related to the reasoning and visual analysis of the signal. We demonstrate the tool's effectiveness by presenting 9 examples with several types of queries on time series. The SSTS queries were compared with standard code developed in Python, in terms of cognitive effort, vocabulary required, code length, volume, interpretation and difficulty metrics based on the Halstead complexity measures. The results demonstrate that this methodology is a valid approach and delivers a new abstraction layer on data analysis of time series.     

Automatic new topic identification using multiple linear regression

The purpose of this study is to provide automatic new topic identification of search engine query logs, and estimate the effect of statistical characteristics of search engine queries on new topic identification. By applying multiple linear regression and multi-factor ANOVA on a sample data log from the Excite search engine, we demonstrated that the statistical characteristics of Web search queries, such as time interval, search pattern and position of a query in a user session, are effective on shifting to a new topic. Multiple linear regression is also a successful tool for estimating topic shifts and continuations. The findings of this study provide statistical proof for the relationship between the non-semantic characteristics of Web search queries and the occurrence of topic shifts and continuations.     

Mix and match: combining terms and operators for successful Web searches

This paper presents a detailed analysis of the structure and components of queries written by experimental participants in a study that manipulated two factors found to affect end-user information retrieval performance: training in Boolean logic and the type of search interface. As reported previously, we found that both Boolean training and the use of an assisted interface improved the participants' ability to find correct responses to information requests. Here, we examine the impact of these training and interface manipulations on the Boolean operators and search terms that comprise the submitted queries. Our analysis shows that both Boolean training and the use of an assisted interface improved the participants' ability to correctly utilize various operators. An unexpected finding is that this training also had a positive impact on term selection. The terms and, to a lesser extent, the operators comprising a query were important factors affecting the participants' performance in query tasks. Our findings demonstrate that even small training interventions can improve the users' search performance and highlight the need for additional information retrieval research into how search interfaces can provide superior support to today's untrained users of the Web.     

Social summarization in collaborative web search

A critical challenge for Web search engines concerns how they present relevant results to searchers. The traditional approach is to produce a ranked list of results with title and summary (snippet) information, and these snippets are usually chosen based on the current query. Snippets play a vital sensemaking role, helping searchers to efficiently make sense of a collection of search results, as well as determine the likely relevance of individual results. Recently researchers have begun to explore how snippets might also be adapted based on searcher preferences as a way to better highlight relevant results to the searcher. In this paper we focus on the role of snippets in collaborative web search and describe a technique for summarizing search results that harnesses the collaborative search behaviour of communities of like-minded searchers to produce snippets that are more focused on the preferences of the searchers. We go on to show how this so-called social summarization technique can generate summaries that are significantly better adapted to searcher preferences and describe a novel personalized search interface that combines result recommendation with social summarization.