to my home pageA high-dimensional semantic space model is a method for numerically representing the meanings of words; it is derived from the frequency distributions of the words occurring in the immediate context of a target word, computed over a large language corpus (containing millions of words). Words that occur in the same sorts of contexts are thus contextually similar, and tend to be similar in meaning.
You can do two things with this interface: (1) calculate the Contextual Similarity between a target word and another word, or between a target word and a set of comparison words; (2) compute the 'nearest neighbours' of a target word. The default similarity measure is the cosine of the angle between word vectors; if you choose the Normalise option, then similarity values are transformed into z-scores, allowing words of markedly different frequency (with consequent differences in vector sparseness) to be legitimately compared. NB: if using any of the lemmatised semantic space models, you need to make sure that your test words are in their base [non-inflected] forms.
Huettig, F., Quinlan, P. T., McDonald, S. A. & Altmann, G. T. M. (2006). Models of high-dimensional semantic space predict language-mediated eye movements in the visual world. Acta Psychologica, 121, 65-80.
McDonald, S. (2000). Environmental determinants of lexical processing effort. PhD dissertation, University of Edinburgh.
McDonald, S. & Lowe, W. (1998). Modelling functional priming and the associative boost. Proceedings of the 20th Annual Conference of the Cognitive Science Society, 675-680. Mahwah, NJ: Erlbaum.
McDonald, S. & Ramscar, M. (2001). Testing the Distributional Hypothesis: The influence of context on judgements of semantic similarity. In Proceedings of the 23rd Annual Conference of the Cognitive Science Society.
Below are links to two web-interfaces that do the same thing, but in slightly different ways. Compare results!
to my home page