The Frame-Based Module of the SUISEKI Information Extraction System

names in the same abstract already indicates a relation between them. Because of its simplicity, we can apply this approach to large amounts of text and establish interaction networks for Saccharomyces cerevisiae and humans.4,5 The limiting factor is that we can't know the type of relation between the genes and proteins this way. To address this problem, we developed SUISEKI, an information extraction system that takes an intermediate view of the problem by requiring the two names to be in a frame that indicates a direct or indirect interaction between them. The SUISEKI system On the one hand, SUISEKI uses statistics and frequency of occurrence, while on the other, it uses analysis of the syntactical structure of phrases and other developments in computational linguistics. This simplified view of the possible complications in the context of text analysis finds its justification in the field of natural language understanding. Boththegrammarandpattern-matchingapproaches offer advant...

• Type: Journal
• Year: 2002
• Where: EXPERT
• Authors: Christian Blaschke, Alfonso Valencia
• Keywords: EXPERT 2002, Interaction Networks, Large Amounts, Protein