Development and evaluation of a prototype system for helping indexers, including authors and publishers, produce traditional back-of-the-book indexes.
We propose to develop and test a prototype system for helping indexers (including authors, scholars, and publishers) produce traditional back-of-the-book indexes. Using standard text analysis technology (including Latent Semantic Analysis, Named Entity Extraction, Hierarchical Cluster Analysis, and other methods) we hope to identify, group, and present appropriate concepts for inclusion in an index and then automatically generate index anchors within the text itself. Human input will be possible -- and indeed, encouraged -- at any point in the process.