This Small Business Innovation Research Phase I project from TextWise, LLC seeks to develop the theoretical basis for a legal datamining system. The "DOCKET" system will be a natural language-based datamining tool that will automatically extract name entities such as people, places and organizations from electronic documents. Extracted information will be presented in the form of concept-relation-concept triples and organized chronologically. Relationships between extracted entities will be displayed graphically. Links to the original documents will allow users to go directly to source documents of interest. Specifically, in Phase I, TextWise will focus on a structural analysis of email messages as a text type. The current near ubiquity of email as a form of intra- and inter-organizational communication has exacerbated the administrative challenge inherent in the legal discovery process-generating many more written documents than was previously the case. Development of a system that performs an automatic analysis of email and similar textual material will result in substantial improvements in the quality and timeliness of pre-trial activities. Information can thus also be extracted from optically scanned transcripts, depositions and interrogatories. The technology proffered by TextWise will contribute to litigation support services such as processing and visual summarization of relevant documents and stand-alone investigative tools such as a detective's electronic case notebook of persons, places and times. The markets for DOCKET are law firms engaged in litigation and state and federal administrative agencies as well as other private firms whose activities affect life, liberty or property.
