Since the multiple attribute intent weight values are slightly bigger than the actual weighted values when weighted concept sub-lattices based on information entropy and deviance being combined

thresholds disturbed by subjective factor directly causes the merged concept lattice size to be exaggerated

an improved optimization algorithm of weighted concept lattices is proposed. Multiple attribute intent weight values are decided by the whole entropy of attributes sets in data sets. Threshold interval of each concept node multiple attribute intent weight value is computed by variance to overcome the subjective factors. The size of the weighted concept lattice construction and time-consuming are reduced by removing redundant nodes which does not satisfy the threshold interval. The experimental results indicate that the proposed algorithm is reduced 9. 87 % redundant nodes

the time-consuming of whole concept lattice construction is decreased by 7. 36 %. The proposed algorithm apparently improves the efficiency of constructing weighted concept lattices.