Research And Application Of Apriori Algorithm In Association Rules

At present,the whole society enters the information age,which is characterized by massive information and data.In the era of big data,data is the intangible wealth and assets of human beings.With the continuous generation of massive data,new technical means and tools must be used to process massive data sets,so as to help extract useful information and rules from the data more intelligently,and mine the “wealth and value” in the data.Among the new technical means and tools,association rule mining is one of the important research topics in data mining.The purpose of association rule mining technology is to discover the connection rules between various data,and its core is the generation of frequent itemsets.First of all,this paper introduces the relevant knowledge of data mining and association rules in detail,and deeply studies the most classic Apriori algorithm for association rule mining.The main idea of this algorithm is to effectively mine the frequent itemsets of the transaction database,with which people can quickly and accurately find association rules.But the Apriori algorithm is still imperfect because the database needs to be scanned for many times,and a large number of useless candidate itemsets are generated during the operation of the algorithm.Secondly,this paper proposes two improved algorithms for the defects of Apriori algorithm: Z＿Apriori algorithm and E＿Apriori algorithm.The idea of Z＿Apriori algorithm is to delete itemsets that do not meet the conditions in advance,reduce frequent candidate sets,and avoid excessive scanning itemsets when scanning the database,thereby improving the efficiency of the algorithm.The idea of the E＿Apriori algorithm is aimed at the unique support of the Apriori algorithm.Each record in the transaction record can be assigned a certain weight to affect the proportion of data in the data set,so that the association rules are affected by other attributes.The rules are biased towards the one with the most weight,and the association rules obtained in this way can be better used in practical applications.Finally,the implementation process of the three algorithms is analyzed in detail through pseudo-code,and the three algorithms are respectively run in the same instance,and the experimental results are analyzed.Greater than the conclusion of the classic Apriori algorithm.The three algorithms are then used in practical applications,the Movie Lens dataset is used and the One-hot encoding method is used for data preprocessing,and the processed dataset is applied to the three algorithms to realize movie recommendation based on movie tags.