Captive balloon fault association rule mining method and system based on data mining

The invention provides a data mining-based captive balloon fault association rule mining method and system. The method comprises the following steps of: counting the fault data frequency of a captive balloon subsystem; creating an initial frequent set L1 with the length of 1; creating a candidate set CK by the supersets Lk-1 of all the frequent sets; judging whether items in the candidate set are greater than the minimum support degree or not; if yes, the frequent item set is reserved; judging whether the confidence coefficient of the frequent item set is greater than the minimum confidence coefficient; if yes, judging the rule as a preliminary strong association rule; if not, determining the weak association rule as a preliminary weak association rule; reading the state data of the key weight of the captive balloon in the fault time period; carrying out regularization processing on abnormal values; carrying out data normalization processing; mutual information method filtering; quantitatively analyzing the c