Our approach evaluates the performance using a mathematical metric. However, it is the human who finally judges the goodness of the generated hierarchy. Therefore, subjective evaluations that are conducted on human subjects are necessary for con-structing a human-readable hierarchy. The result of this work will be shown to several human subjects to evaluate the result in the near future. In addition, the constructed hierarchies are to be used for text categorization. Both subjective and objective evalu-ations of the performance of text categorization will also be conducted. Finally, we will compare the effectiveness of using our hierarchies for text categorization with the hierarchies constructed by some other methods.
| |||||||||||||||||||||||||||||||||
|