Information synthesis: A mixed-initiative meta-analytic approach to facilitate knowledge discovery from scientific text

The quantities of electronic information resources continue to increase at an overwhelming rate. Although retrieval systems have eased the task of collecting articles, users struggle to incorporate new findings into their work practices and have few opportunities to explore implicit connections within a collection of information resources. My goal is to develop methodologies and associated technologies that will facilitate discovery from textual information resources.; Inspired by my study of users in medicine and public health as they synthesized evidence from literature, I introduce a new process called Information Synthesis. I designed the Information Synthesis process for users who make consequential decisions and who operate in an intensive environment. The process increases a user's opportunity to detect novel phenomena by incorporating typically unused information from scientific articles. I developed the M ulti-user Extraction for Information Synthesis ( METIS) system to demonstrate the feasibility of the information synthesis approach. METIS automates critical tasks within the information synthesis process which are to (1) identify information from full-text articles, (2) estimate a comparison group using the extracted information as an index to an external database, and (3) compare the facts reported in each article with the comparison group using a meta-analytic technique. The METIS generated quantitative and visual summaries reflect the redundancies and contradictions that are inevitable in an information intensive environment.; I make three claims regarding the METIS system: (1) the shallow natural language processing used within the METIS information extractor identify information items at a level of precision and recall that are consistent with current state of the art; (2) the METIS comparison estimator estimates a comparison group rate that is similar to the rates found in a traditional analysis; and (3) the METIS analyzer produces the same effect size as published analyses.; I evaluated the information synthesis empirically by exploring risk factors within a corpus of breast cancer articles. My results show that the information synthesis approach was (1) more comprehensive than existing published analyses that explored a similar hypothesis projection, and (2) able to identify phenomena that existing synthesis techniques were unable to detect.