AIMS: Assisted Indexing at Mississippi State

In this report, we describe the development of an automated system that aids human document analysts in the assignment of indexes to journal articles in the area of physical chemistry. AIMS (Assisted Indexing at Mississippi State) has two major components: a natural language processing (NLP) component and an index generation (IG) component. This report provides an overview of each component. The description of the NLP component focuses on the identification of units of the text and the attachment of appropriate tags to terms in the text. Primarily, there are two kinds of units of interest: phrasal units (such as prepositional phrases) and word collocations in which the “multi-word” may differ semantically from any of the individual words that make up the multi-word (e.g., “atmospheric pressure”). The tags include both syntactic and semantic tags (e.g.: noun/chemical substance). The discussion of the IG component describes how this component makes use of the parsed articles produced by the NLP component and some domain-specific information such as a hierarchy of chemistry concepts to generate a set of recommended indexes for a given article.