Research On Hybrid Recommendation System Based On BQ-Apriori Algorithm

With the rapid development of intelligent algorithms, the role of intelligent algorithms in the recommendation system is becoming more and more important. The intelligent recommendation algorithm has become a very active frontier research topic. Compared with the traditional recommendation algorithm, the intelligent recommendation algorithm has the advantages of personalization and precision. Nowadays, the popular "Moobai Bicycle" uses many intelligent recommendation algorithms, such as the intelligent recommended parking point can be accurate and rapid positioning bike. Intelligent recommendation algorithm is also more and more widely used in e-commerce system, but the current intelligent recommendation algorithm is often not the best customers, on the one hand, customers can not quickly and accurately search for their favorite goods; The system recommended to the customer’s goods are often not the customer’s favorite. Aiming at the shortcomings of the proposed algorithm,this paper proposes a kind of intelligent recommendation system solution based on the improved Apriori algorithm. The aim is to provide high quality personalized recommendation service for customers in a timely and accurate way.First of all,in order to achieve the above purpose,this paper proposes a BQ-Apriori algorithm based on transaction deletion based on Apriori algorithm. Based on the algorithm, the corresponding algorithm model is established. The algorithm model is based on real of the agricultural products trading system based on. Agricultural products trading system source a real agricultural product sales system,mainly through the agricultural trading platform to sell high-end agricultural products to meet the needs of the majority of high-end users, making them in the shortest possible time to enjoy the agricultural products to the home. In this paper, the real agricultural product transaction data is used as the reliable data source for the agricultural product transaction recommendation system, and the improved algorithm is verified and analyzed by simulation experiment. It is further proved that the improved algorithm is reliable and accurate in dealing with the related problems.Secondly, according to the basic theory of a priori knowledge, the initial data set of the algorithm is processed, and the basic structure of the algorithm is designed. By using the basic properties of the algorithm, a kind of transaction-based improved new association rule algorithm. With the help of the main idea of Apriori algorithm, the basic idea of Apriori algorithm is to find the process of the largest frequent itemsets. The second frequent itemsets are searched by the first frequent itemsets and the third frequent itemsets are searched by the second frequent itemsets. Analogy, and finally get the largest frequent itemsets, and then through the largest frequent itemsets to find strong association rules, through its in-depth understanding of the idea of the algorithm to delete the results of the implementation of the algorithm did not affect the single number of transactions, followed by a priori the knowledge constructs the set of customers interested, sets the transaction items related to the collection of interests as the research points, and eliminates the transaction items that are irrelevant to the result. As a result, the number of database sets scanned by the algorithm is greatly reduced,which makes the time complexity of the algorithm greatly reduced.Finally, the application of the algorithm and the analysis of the model. In this paper, the data in the real agricultural trade system is taken as the experimental data, and the concrete realization of the BQ-Apriori algorithm based on transaction deletion is discussed. The association information of the commodity information of the agricultural product trading system is analyzed, regular relationship. Then when the customer orders, will generate an order in the database, we put the associated goods on the customer orders behind, marked the purchase of this product customers also bought other goods. Then, the basic idea of SOM neural network is studied, and the membership level of SOM neural network is classified by SOM neural network. The members can be classified according to the classification of the members,and the results of the classification and the results of commodity association analysis applied to the hybrid recommendation system to go. Eventually, a hybrid recommended system model was established.