Abstract
A dynamic fuzzy rule promotion approach for the promotion of a confidence factor of a rule for every successful session in diagnosis of a disease in crops by using the specific rules, has already been proposed in literature. This technique has the limitation that an error in the initial estimation of weights reduces linearly after every session the rule is being used. In this paper an improved approach has been proposed using the square root of sum of squares of frequencies, which are spread around the mean true value to reduce the error around a mean value. A rule set for the diseases and their symptoms for the paddy plant has been provided to make comparison between the previous and the improved approach. It has been shown that the improved approach decreases the error in uncertainty of estimation of weight for rules after every successful session. It has also been proposed that the improved approach must be applied in agricultural information dissemination system.