Jon Ducrou

Concept Lattices for Information Visualization: Can Novices Read Line-Diagrams?

Mail-Sleuth is a personal productivity tool that allows individuals to manage email and visualize its contents using line diagrams. Based on earlier work on the Conceptual Email Manager (Cem), a major hypothesis of Mail-Sleuth is that novices to Formal Concept Analysis can read a lattice diagram. Since there is no empirical evidence for this in the Formal Concept Analysis literature this paper is a first attempt to test this hypothesis by following a user-centred design and evaluation process. Our results suggest that, with some adjustments, novice users can read line diagrams without specialized background in mathematics or computer science. This paper describes the process and outcomes based on usability testing and explains the evolution of the Mail-Sleuth design responding to the evaluation at the Access Testing Centre.

FCA-based browsing and searching of a collection of images

This paper introduces ImageSleuth, a tool for browsing and searching annotated collections of images. It combines the methods of Formal Concept Analysis (FCA) for information retrieval with the graphical information conveyed in thumbnails. In order to use thumbnails of images to represent concept extents, line diagrams can not be efficiently utilised and thus other navigation methods are necessary. In addition to established methods like search and upper/lower neighbours, a query by example function and the possibility to restrict the attribute set are included. Moreover, metrics on conceptual distance and similarity are discussed and applied to automated discovery of relevant concepts. This paper describes the FCA base of ImageSleuth which formed the basis for its design and the implementation which followed.

An intelligent user interface for browsing and searching MPEG-7 images using concept lattices

This paper presents the evaluation of a design and architecture for browsing and searching MPEG-7 images. Our approach is novel in that it exploits concept lattices for the representation and navigation of image content. Several concept lattices provide the foundation for the system (called IMAGE-SLEUTH) each representing a different search context, one for image shape, another for color and luminance, and a third for semantic content. This division of information aids in the facilitation of image browsing based on a metadata ontology. The test collection used for our study is a sub-set of MPEG-7 images created from the popular The Sims 2™ game. The evaluation of the IMAGE-SLEUTH program is based on usability testing among 29 subjects. The results of the study are used to build an improved second generation program - IMAGE-SLEUTH2- but in themselves indicate that image navigation via a concept lattice is a highly successful interface paradigm. Our results provide general insights for interface design using concept lattices that will be of interest to any applied research and development using concept lattices.

Concept Similarity and Related Categories in SearchSleuth

SearchSleuth is a program developed to experiment with the automated local analysis of Web search using formal concept analysis. SearchSleuth extends a standard search interface to include a conceptual neighborhood centered on a formal concept derived from the initial query. This neighborhood of the concept derived from the search terms is decorated with its upper and lower neighbors representing more general and specialized concepts respectively. In SearchSleuth, the notion of related categories --- which are themselves formal concepts --- is also introduced. This allows the retrieval focus to shift to a new formal concept called a sibling. This movement across the concept lattice needs to relate one formal concept to another in a principled way. This paper presents the issues concerning exploring and ordering the space of related categories.

DVDSleuth: A Case Study in Applied Formal Concept Analysis for Navigating Web Catalogs

Browsing images using Formal Concept Analysis (FCA) for conceptual representation, navigation and clustering was shown in the ImageSleuth projects [1,2,3]. To showcase the ideas and knowledge gained through ImageSleuth, the same techniques were applied to an information space built from a dynamic collection sourced from the Amazon.com on-line store. Using the Amazon.com catalog, conceptually similar DVDs are able to be discovered and viewed, and then used to explore the information space of their conceptual neighbourhood. A case study of the project --- called DVDSleuth --- is presented in this paper focusing on the history, results and difficulties encountered by the project. The shortcomings of our approach are analysed and reported as a guide for future projects using FCA techniques for information exploration using on-line Web catalogs.

AN INTELLIGENT USER INTERFACE FOR BROWSING AND SEARCHING MPEG-7 IMAGES USING CONCEPT LATTICES

This paper presents the evaluation of a design and architecture for browsing and searching MPEG-7 images. Our approach is novel in that it exploits concept lattices for the representation and navigation of image content. Several concept lattices provide the foundation for the system (called IMAGE-SLEUTH) each representing a different search context, one for image shape, another for color and luminance, and a third for semantic content, namely image browsing based on a metadata ontology. The test collection used for our study is a sub-set of MPEG-7 images created from the popular The Sims 2™ game. The evaluation of the IMAGE-SLEUTH program is based on usability testing among 29 subjects. The results of the study are used to build an improved second generation program – IMAGE-SLEUTH2 – however these results also indicate that image navigation via a concept lattice is a highly successful interface paradigm. Our results provide general insights for interface design using concept lattices that will be of interest to any applied research and development using concept lattices.

Concept similarity and related categories in information retrieval using formal concept analysis

The application of formal concept analysis to the problem of information retrieval has been shown useful but has lacked any real analysis of the idea of relevance ranking of search results. SearchSleuth is a program developed to experiment with the automated local analysis of Web search using formal concept analysis. SearchSleuth extends a standard search interface to include a conceptual neighbourhood centred on a formal concept derived from the initial query. This neighbourhood of the concept derived from the search terms is decorated with its upper and lower neighbours representing more general and special concepts, respectively. SearchSleuth is in many ways an archetype of search engines based on formal concept analysis with some novel features. In SearchSleuth, the notion of related categories – which are themselves formal concepts – is also introduced. This allows the retrieval focus to shift to a new formal concept called a sibling. This movement across the concept lattice needs to relate one formal concept to another in a principled way. This paper presents the issues concerning exploring, searching, and ordering the space of related categories. The focus is on understanding the use and meaning of proximity and semantic distance in the context of information retrieval using formal concept analysis.

Web services and digital ecosystem support using formal concept analysis

This paper describes a Web Services (WS) and distributed systems architecture for Formal Concept Analysis that supports information and content management in a social media system. As well as social tagging, the system includes novel approaches to document browsing and heterogeneous Web information retrieval. The Formal Concept Analysis WS architecture supports a system called the Virtual Museum of the Pacific, a content and knowledge acquisition tool that permits the machine synthesis of formal concepts and provides a Rich Internet Application in which to display and navigate them. The interface also allows the extensible association of digital objects via introducing new attributes and relationships. The WS architecture and user interface form the basis for a novel distributed content management and social media application and is to our knowledge the first implementation of Formal Concept Analysis as WS.

Combining spatial and lattice-based information landscapes

In this paper we report on practical information visualization aspects of Conceptual Knowledge Processing (CKP), realizing and illustrating Willes “conceptual landscapes” in the context of developing a conceptual information system to determine surfing conditions on the South Coast of New South Wales, Australia. This novel application illustrates some (if not all) of Willes CKP tasks: exploring, searching, recognizing, identifying, analyzing, investigating, deciding, restructuring and memorizing (all but improving). It does this by concentrating on combining an information landscape with maps of the physical world.

D-SIFT: a dynamic simple intuitive FCA tool

This paper introduces D-SIFT, a Web-based browser application that provides untrained users in Formal Concept Analysis with practical and intuitive access to core analysis functionality in Formal Concept Analysis. D-SIFT is an information systems architecture that supports natural search processes over a predefined database schema and its attribute values. This enables the user to build concept lattices interactively through the selection and refinement of dynamic definitions of search boundaries (via interaction with an object “zoom” feature), and dynamic selection of search scales (via interaction with an attribute “filter” feature), based on the attribute values contained within the database. The paper investigates the claim that D-SIFT systems are an advance on the search and analysis paradigm of the Toscana-system workflow. In detail, the paper presents the architecture of the D-SIFT browser and illustrates the resulting D-SIFT-systems on an example database. The two examples illustrate the generality of system integration outcomes from D-SIFT. The Conceptual Information Systems that result from applying the D-SIFT architecture present a new workflow for building and interacting with Formal Concept Analysis-based information systems. The workflow more closely aligns with dynamic schema interaction, a popular technique used in conceptual modeling.