Keyword-based Text Matching Approach for Design Style Recognition

摘要

We present the results of an investigation into the recognition a design style by analysing keywords in the text descriptions of design styles. A simple keyword-based matching technique is used to classify a design style by examining its text description. Domain specific dictionaries of keywords are used to reduce the dimensions of the feature space. The results of the classifier are compared with those of SVM and decision tree based classifiers. The results conclude that design style in the domain that we analysed can be recognised with accuracy of approximately 75% from its descriptions.