Methods and Software for Mining Data in Context
TBMG-7009
09/01/2002
- Content
The Perilog methods and software provide a suite of data-mining tools. Perilog retrieves and organizes data from any sequence of terms that are contextually associated (e.g., text, music, genetic sequences). Perilog data mining includes: (1) keyterm-in-context search; (2) flexible, model-based phrase search; (3) model-based phrase generation; and (4) narrative-based phrase discovery. The methods and software measure the degree of contextual association for large numbers of paired terms to produce models that capture the structures of the sequences of terms. Perilog compares these models to a query model, develops a ranking, and presents the results of the search to the user. It can also extract phrases from the sequences. Perilog was originally designed to support the Aviation Safety Reporting System, which NASA operates for the Federal Aviation Administration. In that application it produced, from a database of tens of thousands of documents, the first quantitative evidence of situational relationships between reported commercial aviation incidents and a specific type of aviation accident. Perilog can be used to mine databases for any type of information.
- Citation
- "Methods and Software for Mining Data in Context," Mobility Engineering, September 1, 2002.