Research On Association Rule Data Mining Algorithms For Plant Information Detection

With the rapid development and wide application of science and technology in information,the application scale of information systems in various industries is constantly expanding,so the amount of information shows an explosive growth trend,then big data emerges from this.Under this background,using relevant technical methods enables users to find interesting and valuable information from a large number of data sets in a short time,which is one of the important directions for the development of big data processing in the future.With the rapid increase in the amount of data,the number of association rules generated by data mining continues to increase,and the useless redundant association rules continue to increase,which enables users to find valuable rules and increases the difficulty of understanding association rules.And applications and decision-making both cause a lot of difficulties.This paper firstly aims at improving the shortcomings of Apriori algorithm,and establishing a DAD operator model based on Apriori algorithm.By collecting the spectral reflection values of strawberry leaves under different water conditions,two indexes of correlation coefficient and error rate were established,and the proposed operator model was used to correlate the potential relationship between the water content of canopy leaves and the spectral reflectance at different wavelength.The rule analysis found the correlation between the spectral reflection characteristics of strawberry and the leaf water condition under four different water treatment methods:drought,mild drought,moderate and overflow.Secondly,for the criterion of deleting redundant association rules,three inspection indexes are introduced,and the redundant association rules are deleted with minimum support and minimum confidence.The weight analysis of the three test indexes is demonstrated in detail.The experimental results show that the three test indexes can effectively delete redundant association rules.Third,when performing K-means clustering analysis,the maximum triangle method is used instead of the original method which is randomly selecting initial points,and the association rules are clustered to make similar association rules fall into the same cluster.The algorithm is verified by using the iris data set.Experiments show that the improved algorithm not only improves the performance of clustering,but also increases the user's understanding of association rules,which helps users analyze the rules and make correct decisions.Fourth,this article visualizes the association rules and adds dynamic display of rules.Users can select and filter rules through external devices.The selected rules will display detailed information on the left console,at the same time improving users' understanding of rules and enhancing users' experience.