Joris D'hondt

Identifying candidates for design-by-analogy

Although supported by extensive anecdotal evidence, only recently design-by-analogy has been proven to occur often in problem solving and idea generation. However, the circumstances which facilitate problem solving and creative idea generation by analogy are not well understood and most analogies are not developed by applying a formal design-by-analogy methodology. Furthermore, most software tools which aid in finding and/or explaining analogies are based on manually assembled databases, which require a large amount of interactive work to be constructed and maintained. This paper examines the use of automatically distilled product characteristics, called Product Aspects, as a way to automatically and systematically identify candidate products for design-by-analogy. Case studies illustrate this idea generation methodology for three different target products.

Pairwise-adaptive dissimilarity measure for document clustering

This paper introduces a novel pairwise-adaptive dissimilarity measure for large high dimensional document datasets that improves the unsupervised clustering quality and speed compared to the original cosine dissimilarity measure. This measure dynamically selects a number of important features of the compared pair of document vectors. Two approaches for selecting the number of features in the application of the measure are discussed. The proposed feature selection process makes this dissimilarity measure especially applicable in large, high dimensional document collections. Its performance is validated on several test sets originating from standardized datasets. The dissimilarity measure is compared to the well-known cosine dissimilarity measure using the average F-measures of the hierarchical agglomerative clustering result. This new dissimilarity measure results in an improved clustering result obtained with a lower required computational time.

Identifying Document Metadata Based on Multilayer Clustering

This paper presents a novel technique to semi-automatically identify metadata for documents when installing a knowledge management system. Document management systems often deal with large collections of documents. This vast amount of information needs to be searchable for the knowledge worker. Supporting techniques are needed to aid the knowledge worker in his search for information. Many of these techniques are based on the presence of metadata for each document. The techniques presented in this paper are based on a novel approach called multilayer clustering. Using this clustering technique, documents can be assigned to one or more document types. Besides this assignment to a specific type, properties and values are assigned to this document based on term networks extracted from this document. The preliminary tests presented in this paper were performed on a public and several private dataset. The results obtained from the tests indicate that this approach is promising.