Language models for hierarchical summarization
Hierarchies have long been used for organization, summarization, and access to information. In this dissertation we define summarization in terms of a probabilistic language model and use this definition to explore a new technique for automatically generating topic hierarchies. We use the language model to characterize the documents that will be summarized and then apply a graph-theoretic algorithm to determine the best topic words for the hierarchical summary. This work is very different from previous attempts to generate topic hierarchies because it relies on statistical analysis and language modeling to identify descriptive words for a document and organize the words in a hierarchical structure.
We compare our new technique to previous methods proposed for constructing topic hierarchies, including subsumption and lexical hierarchies. We also compare the words chosen to be part of the hierarchy to the top ranked words using TF.IDF in terms of how well each summarizes the document set. Our results show that the language modeling approach performs as well as or better than these other techniques in non user-based evaluations. We also show that the hierarchies provide better access to the documents described in the summary than does a ranked list using one of the non-user based evaluations we have developed. In a user study that compares the ability of users to find relevant instances using both the hierarchy and a ranked list to using the ranked list alone, we find that users like the information provided by the hierarchy and after some practice can use it as effectively as they can a ranked list.