Research Seminar
Determining the complexity of semantic category representationsMatt Dry, Steven Verheyen, & Gert StormsKU Leuven | |
| Abstract: | Multidimensional scaling (MDS) is a widely used method of representing semantic category structure. A pervasive problem in this area relates to determining the optimal dimensionality of the representation. At one extreme it is possible to underestimate the number of dimensions in an attempt to provide a simple, visually interpretable solution. At the other extreme it is possible to overestimate the number of dimensions in an attempt to account for as much of the empirical variance as possible. Verheyen, Ameel & Storms (in press) reviewed a number of techniques for determining optimal dimensionality and introduced a new method based upon an external empirical criterion. One of the advantages of such an approach is that any improvement in fit that is obtained by increasing the dimensionality of the representation cannot be simply due to fitting noise in the empirical input data. However, the approach outlined in Verheyen et al. is unable to provide an indication of the cost (in regards to model complexity) of any improvement in fit. In this talk we outline a Bayesian reworking of the Verheyen et al. approach that is inherently sensitive to the competing pressures of goodness-of-fit and model complexity. |
| Date: | Tue May 22, 12:15 pm - 1:15 pm |
| Place: | room 00.60 (Department of Psychology, Tiensestraat 102, 3000 Leuven) |