Abstract: In this paper the problem of self-learning in the Post-earthquake Tendency Decision Support System (PTDSS) is discussed. The method of Fussy Associative Memory is introduced for solving this problem, and a digital method for determination of earthquake sequence type is presented. A great deal of meaningful experience are obtained in this system by self-learning which can be used to determine the sequence type in early stage of sequence. The examples show that this digital method is more efficient for the classification of sequence type than the traditional one.

Key words: Earthquake sequence, Earthquake type, Machine learning, Neural network, Fussy associative memory FAM model