Accession Number : ADA314958

Title :   Detection and Classification of Signals and Noise with Long Memory.

Descriptive Note : Final rept. 1 Dec 89-31 Jan 94,

Corporate Author : BOSTON UNIV MA DEPT OF MATHEMATICS

Personal Author(s) : Taqqu, Murad S. ; Samorodnitsky, Gennady

PDF Url : ADA314958

Report Date : 29 JAN 1996

Pagination or Media Count : 16

Abstract : Long memory occurs when low frequencies have a fundamental impact on the dependence structure of the data. There may also be high variability which occurs when the data has fat distribution tails. Methods were developed for the detection and classification of signals with such characteristics. Some of these techniques were applied to the analysis of computer traffic. The corresponding article, authored by Leland, Taqqu, Willinger and Wilson, was reprinted a number of times. Its extended version has received the 1995 William J. Bennett Award from the IEEE Communications Society and the 1996 IEEE W.R.G. Baker Prize Award. The Baker Prize Award recognizes 'the most outstanding paper reporting original work' in all publications of the IEEE.

Descriptors :   *MATHEMATICAL MODELS, *SIGNAL TO NOISE RATIO, *COMPUTER NETWORKS, DATA BASES, SIGNAL PROCESSING, OPTIMIZATION, STOCHASTIC PROCESSES, QUEUEING THEORY, DATA MANAGEMENT, DISTRIBUTED DATA PROCESSING, COMPUTER COMMUNICATIONS, COMMUNICATIONS TRAFFIC, TIME DEPENDENCE, RANDOM VARIABLES, STATISTICAL INFERENCE, TIME SERIES ANALYSIS, ANALYSIS OF VARIANCE, GAUSSIAN NOISE, LONG RANGE(TIME), REGRESSION ANALYSIS, SYSTEMS ANALYSIS, BROWNIAN MOTION.

Subject Categories : Computer Systems
      Operations Research

Distribution Statement : APPROVED FOR PUBLIC RELEASE