الفهرس Only 14 pages are availabe for public view
 |  | from | 138 |  |  |
| | | from | 138 | | |
Abstract
TIME SERIES ANALYSIS PLAYS AN IMPOTTANT ROLE IN MANY ASPECTS OF DIFFERENT FIELDS OF APPLICATION. USUALLY REEI PROCESSES WHICH GENERATE TIME SERIES HAVE SOME FORM OF NONLINEARITIES. HOWEVER, FOR ANALYSIS, A LINEAR MODEL IS PREFFERED FOR THE GENERATING PROCESS INorder TO REDUCE COMPUTATION ENVOLVED. THE ACCURACY OF THE CALCULATED LINEAR MODEL IS IMPORTANT ANDDEPENDS ON THE CHOSE STRUCTURE. IN CASES where COEFFICIENTS OF THIS MODEL ARE NOT KNOWN, CHARACTERISTICS OF GENERATING PROCESS CAN ALWAYS BE EXTRACTED from INPUT OUTPUT DATA, AND THIS DAEA MAY BE COLLECTED BY RANDOM SIGNAL TESTING. IN THIS CASE, POWER OF TEST SIGNAL IS REQUIRED TO COVER MOST DYNAMICAL MODES IN THE GENERATING PROCESS. THE ACCURACY OF THE RESULTED MODEL DEPENDS ON THE NUMBER OF DATA POINTS OR PREDICT OR ESTIMATE THE FUTURE BEHAVIOUR OF THE GENERATING PROCESS IN TERMS OF DATA UP TO PRESENT.