Abstract: A new approach to trouble pattern analysis is proposed for after service of products based on data mining.Dimensionality of the fault information is collected, and the fault is cleaned up.The dimensionality table of association rules is made up of technical parameters and the causes of fault.The fault information is analyzed by a set of frequent items with the apriori algorithm based on the dimensionality tables of fault information and association rules.Causes of fault are found and the primary solution is chosen by fault matching, candidate generation and candidate screening.

Key words: trouble pattern analyze, data mining, frequent item set, trouble match