Free-Text Search over Complex Web Forms

Abstract

Abstract. This paper investigates the problem of using free-text queries as an alternative means for searching`behind' web forms. We introduce a novel speci cation language for specifying free-text interfaces, and report the results of a user study where we evaluated our prototype in a travel planner scenario. Our results show that users prefer this free-text interface over the original web form and that they are about 9% faster on average at completing their search tasks.Keywords: query processing, free-text interfaces, query translation. IntroductionThe internet contains a large amount of information that is only accessible through complex web forms. Journey planners, real estate websites, online auction and shopping websites, and other websites commonly require the user to ll out a form consisting of a number of elds in a graphical interface. The user should rst interpret the form and then translate his information need to the appropriate elds. Filling out these forms can be slow because they require mixed interaction with both the mouse and keyboard. A natural language interface (NLI) alleviates these problems by allowing the user to enter his information need in a single textual statement. Rather than navigating between and entering information in the components of the web form, the user can focus on formulating his information need in an intuitive way. NLIs require or assume syntactically well-formed sentences as input, in essence restricting the range of textual input.However, describing all possible natural language statements and dealing with query ambiguity can be a time-consuming process [1 4]. Therefore, we introduce a free-text interface (FTI) which allows the user to freely input text without any restrictions. In this paper, we describe and evaluate a prototype system for specifying FTIs to access information behind complex web forms. The system has been designed to specify exible FTIs with relatively little e ort. This work is a stepping stone for further investigation of a single textual interface to access the deep web [5]. Ideally, we wish to use these techniques to build a distributed search system which can search multiple resources, including information behind complex web forms, simultaneously. The contributions of this paper are as follows: i) we introduce a speci cation language for describing free-text interfaces (FTIs) to complex web forms; ii) as a proof of concept, we show that this language can be e ectively used to describe a exible FTI to a travel planner web form; and iii) we demonstrate that users can search faster with an FTI than

An analysis of free-text queries for a multi-field web form

“…Another remarkable observation is that the most likely valid-query template (date time from station to station) corresponds to the template of the example query shown in the Treinplanner interface. In contrast, in an earlier laboratory experiment where users were given different search tasks in various descriptions, the majority of the users did not follow the templates of the descriptions of the search tasks [18]. One way to mitigate this effect is to refrain from showing any example query in the first place, but that would increase the effect of the habitability problem.…”

“…Treinplanner adopts a pattern-based approach to find the best interpretation of the queries. This approach has been described in [18]. A valid query results in one or more interpretations, i.e.…”