Design And Analysis Of Educational Research System Based On Association Rules

Along with the development of computer technology and the increasing popularity of network, people are facing exponential data growth. While relying solely on traditional data analysis methods can not meet the needs of modern humans, in this case, the data mining technology emerges. Data mining can be through large volumes of data mining internal hidden information, unknown forecasting and decision support to help. Mining association rules which is one of the most prosperous areas of data mining technology is currently, and the technique is mainly used to find the relationship between the data, which has positive significance in the contemporary.For association rule mining, in mining methods academia the most classic algorithm is apriori algorithm. However, the disadvantage of this algorithm is very obvious, then its main drawback is that each item is calculated a set of support, you need to scan the database once. This will cause a great degree of I / O overhead, resulting in very poor performance of the algorithm. Accordingly, academic scholars and researchers have conducted a variety of improved algorithms. Chief among them is aprioriTid algorithm and FP-Growth algorithm. aprioriTid algorithm only needs to scan the database once they can complete the mining; FP-Growth algorithm is mainly needed for the mining process information is stored in a tree, and facilitate association rule mining. However, these two important improvements algorithm has its fatal flaw, when the amount of data is large, both to improve the performance of the algorithm is also lower than the performance of apriori algorithm. This paper summarizes the advantages and disadvantages of various improvements algorithm, and based on bit operation, and divide and rule. The improved process to establish a data structure, and this data structure combines the use of arrays and linked lists. And though access bitwise to conduct information storage. When calculated the support, and binary arithmetic operations. So only need to scan database once, it be able to complete the association rule mining operations though a very high speed. In conclusion, the experimental data demonstrates that the improved algorithm did achieve improved performance.Conducting preliminary data mining, the data is stored over the years of the school dean systems for data extraction, cleansing, integration, integration, and establish a shared data center, then using the improved association rules algorithm and hadoop technology in senate analysis system in this paper, the purpose of distributed data processing is achieve faster processing. Finally dig out the relationship between related courses. Related conversion, according to the situation of certain courses, in order to alert the relevant curriculum results by association rules between courses, and to provide early warning information to the relevant students, to achieve urge students to study subjects related, ultimately enable students to get rid of repetition, credit risk of failure to meet the standards.