A FUNDAMENTAL STUDY ON RASIN GROUTING JUDGEMENT PROCEDURE BASED ON LOCAL OUTLIER FACTOR METHOD IN RC MEMBER REPAIRING

Abstract

In this paper, a judgement procedure based on the local outlier factor method which is one of machine learning techniques fed hitting sound spectrums as input data is proposed to judge success or failure of rasin grouting in reinforced concrete member repaired by the inside pressure hardening. the validity of this procedure has been confirmed an insight experiment evaluating the judgement of success and failure of grouting in a reinforced concrete slab member. In addition, effects of number of teacher data used in the procedure on the true positive ratio, true negative ratio and other evaluating indices are disscussed with effect of threshold value employed in this procedure.