The Research Of Data Mining In Optical Fiber Failure Early Warning System

This thesis studies the deficiencies of the classical association rule mining algorithm---Apriori based on the data mining technology and the association rule mining theory, and optimizes Apriori with the application of Hash Function, group tables and the chech of frequent set with the table, and trims the transaction with the giving variable, gives AprioriGroup algorithm. Based on the VS2003 platform, the data mining system of the optical fiber fault early warning is developed. This thesis finds out the frequent set of optical power alarm information in the database based on analyzing mass optical power alarm history in curve, mines the association between the optical power alarm information and the optical fiber running situation, advances the association rule which can be regarded for determination, and achieves the combination of theory and practice.