Wolfgang Holzinger

A versatile model for web page representation, information extraction and content re-packaging

On todays Web, designers take huge efforts to create visually rich websites that boast a magnitude of interactive elements. Contrarily, most web information extraction (WIE) algorithms are still based on attributed tree methods which struggle to deal with this complexity. In this paper, we introduce a versatile model to represent web documents. The model is based on gestalt theory principles---trying to capture the most important aspects in a formally exact way. It (i) represents and unifies access to visual layout, content and functional aspects; (ii) is implemented with semantic web techniques that can be leveraged for i.e. automatic reasoning. Considering the visual appearance of a web page, we view it as a collection of gestalt figures---based on gestalt primitives---each representing a specific design pattern, be it navigation menus or news articles. Based on this model, we introduce our WIE methodology, a re-engineering process involving design patterns, statistical distributions and text content properties. The complete framework consists of the UOM model, which formalizes the mentioned components, and the MANM layer that hints on structure and serialization, providing document re-packaging foundations. Finally, we discuss how we have applied and evaluated our model in the area of web accessibility.

A unified ontology-based web page model for improving accessibility

Fast technological advancements and little compliance with accessibility standards by Web page authors pose serious obstacles to the Web experience of the blind user. We propose a unified Web document model that enables us to create a richer browsing experience and improved navigability for blind users. The model provides an integrated view on all aspects of a Web page and is leveraged to create a multi-axial user interface.

Modelling web navigation with the user in mind

In this research we introduce the ABBA framework for the generation of advanced screen readers. Current solutions do not enable blind users to participate in the Web in a satisfactory way because they are not time-efficient, cumbersome to use, and do not provide enough overview and orientation within a document. Also, they hide away important layout information from the user. Our approach overcomes these limitations by unifying different semantic views of a document into one multi-axial model and making them accessible to the user in an intuitive interface. The benefit to the user is a higher level of control in different navigation scenarios.

Web object identification for web automation and meta-search

Web object identification plays an important role in research fields such as information extraction, web automation, and web form understanding for building meta-search engines. In contrast to other works, we approach this problem by analyzing various spatial, visual, functional and textual characteristics of web pages. We compute 49 unique features for all visible web page elements, which are then applied to machine learning classifiers in order to identify similar elements on other previously unexamined web pages. We evaluate our approach with different scenarios by analyzing the relevance of the chosen features and the classification rate of the applied classifiers. These scenarios focus on understanding search forms from the transportation domain, particularly flight, train, and bus connections. The results of the evaluation are very promising.

A flight meta-search engine with metamorph

We demonstrate a flight meta-search engine that is based on the Metamorph framework. Metamorph provides mechanisms to model web forms together with the interactions which are needed to fulfil a request, and can generate interaction sequences that pose queries using these web forms and collect the results. In this paper, we discuss an interesting new feature that makes use of the forms themselves as an information source. We show how data can be extracted from web forms (rather than the data behind web forms) to generate a graph of flight connections between cities. The flight connection graph allows us to vastly reduce the number of queries that the engine sends to airline websites in the most interesting search scenarios; those that involve the controversial practice of creative ticketing, in which agencies attempt to find lower price fares by using more than one airline for a journey. We describe a system which attains data from a number of websites to identify promising routes and prune the search tree. Heuristics that make use of geographical information and an estimation of cost based on historical data are employed. The results are then made available to improve the quality of future search requests.

Feature-based object identification for web automation

In this paper, we address automatic identification of common functional structures on web pages, a fundamental problem for web automation applications and graphical user interface testing. In contrast to other approaches, we aim to identify relevant patterns without relying on the source code of a web page or keywords, utilizing mostly geometrical and visually perceptible properties. We achieve this by transforming pages into an independent geometrical representation, on top of which we extract a set of features that allows us to employ traditional machine learning techniques for the identification task. We evaluate this approach by analyzing three typical scenarios, reviewing the obtained information retrieval key metrics and estimating the relevance of the chosen features. Our initial results demonstrate the feasibility of the proposed approach.

Exploiting semantic web technologies to model web form interactions

Form mapping is the key problem that needs to be solved in order to get access to the hidden web. Currently available solutions for fully automatic mapping are not ready for commercial meta-search engines, which still have to rely on hand crafted code and are hard to maintain. We believe that a thorough formal description of the problem with semantic web technologies provides a promising perspective to develop a new class of vertical search engines that is more robust and easier to maintain than existing solutions. In this paper, instead of trying to tackle the mapping problem, we model the interaction necessary to fill out a web form. First, during a user-assisted phase, the connection from the visible elements on the form to the domain concepts is established. Then, with help from background knowledge about the possible interaction steps, a plan for filling out the form is derived.

Automated Ontology-Driven Metasearch Generation with Metamorph

We present Metamorph, a system and framework for generating vertical deep Web search engines in a knowledge-based way. The approach enables the separation between the roles of a higher skilled ontology engineer and a less skilled service engineer, which adds new web sources in an intuitive, semi-automatic manner using the proven Lixto suite. One part of the framework is the understanding process for complex web search forms, and the generation of an ontological representation of each form and its intrinsic run-time dependencies. Based on these representations, a unified meta form and matchings from the meta form to the individual search forms and vice versa are created, taking into account different form element types, contents and labels. We discuss several aspects of the Metamorph ontology, which focuses especially on the interaction semantics of web forms, and give a short account of our semi-automatic tagging system.