MU students sort through information overload
Three MU computer science students are receiving national recognition for successfully tackling a segment of today’s information overload dilemma.
Engineering graduate students Wannapa Mahamaneerat, Jason Green and Tetsuya Kobayashi have earned kudos from the American Medical Information Association (AMIA) competition for a computer program they developed that categorizes huge amounts of data so that useful patterns can be inferred from it. The Mizzou Engineering team is one of four international teams to earn the right to present its work at AMIA’s annual symposium on Nov. 13 in Chicago, Ill.
“In a short period, they produced unique software and a Web-based utility for describing health care data. They analyzed a data set of 8 million transactions,” said Kirk T. Phillips, chair of the AMIA Knowledge Discovery and Data Mining Working Group and a biostatistician for the Iowa Health System.
AMIA, a Maryland -based nonprofit organization dedicated to developing and applying informatics to patient care and public health, cosponsored the competition for the first time this year with the federal government’s Agency for Healthcare Research and Quality, Phillips said. All competing teams were required to submit a paper reporting characteristics or patterns they uncovered by electronically analyzing nearly 8 million medical records for significant information.
The Mizzou Engineering students used their “DCMiner” program to sort the information in those records into 15 categories, or domain-concepts. From there, DCMiner allows analysts to quickly explore possible relationships between those categories and presents both written and graphic results.
Mahamaneerat said the team’s winnowing process not only speeds up analysis but paves the way to knowledge that would otherwise be hidden by the data’s sheer bulk. Important health care information may go unnoticed due to the complexity and volume of the information available, the Mizzou Engineering team’s paper said.
“Our tool is trying to make information seeking more efficient, easier and with cleaner results,” said computer science Associate Professor Chi-Ren Shyu, the Mizzou team’s adviser.