Skip to main content
U.S. Department of Energy
Office of Scientific and Technical Information

Knowledge fusion: Comparison of fuzzy curve smoothers to statistically motivated curve smoothers

Technical Report ·
DOI:https://doi.org/10.2172/215312· OSTI ID:215312

This report describes work during FY 95 that was sponsored by the Department of Energy, Office of Nonproliferation and National Security (NN) Knowledge Fusion (KF) Project. The project team selected satellite sensor data to use as the one main example to which its analysis algorithms would be applied. The specific sensor-fusion problem has many generic features, which make it a worthwhile problem to attempt to solve in a general way. The generic problem is to recognize events of interest from multiple time series that define a possibly noisy background. By implementing a suite of time series modeling and forecasting methods and using well-chosen alarm criteria, we reduce the number of false alarms. We then further reduce the number of false alarms by analyzing all suspicious sections of data, as judged by the alarm criteria, with pattern recognition methods. This report gives a detailed comparison of two of the forecasting methods (fuzzy forecaster and statistically motivated curve smoothers as forecasters). The two methods are compared on five simulated and five real data sets. One of the five real data sets is satellite sensor data. The conclusion is the statistically motivated curve smoother is superior on simulated data of the type we studied. The statistically motivated method is also superior on most real data. In defense of the fuzzy-logic motivated methods, we point out that fuzzy-logic methods were never intended to compete with statistical methods on numeric data. Fuzzy logic was developed to handle real-world situations where either real data was not available or was supplemented with either ``expert opinion`` or some sort of linguistic information.

Research Organization:
Los Alamos National Lab., NM (United States)
Sponsoring Organization:
USDOE, Washington, DC (United States)
DOE Contract Number:
W-7405-ENG-36
OSTI ID:
215312
Report Number(s):
LA--13076-MS; ON: DE96007846
Country of Publication:
United States
Language:
English

Similar Records

Knowledge fusion: An approach to time series model selection followed by pattern recognition
Technical Report · Thu Feb 29 23:00:00 EST 1996 · OSTI ID:219426

Knowledge fusion: Time series modeling followed by pattern recognition applied to unusual sections of background data
Technical Report · Thu Feb 29 23:00:00 EST 1996 · OSTI ID:215313

Mining Building Energy Management System Data Using Fuzzy Anomaly Detection and Linguistic Descriptions
Journal Article · Fri Aug 01 00:00:00 EDT 2014 · Transactions on Industrial Informatics · OSTI ID:1156912