Improved FP-Tree Based Algorithm For Adaptive Learning System In The Characteristics Of Learners In The Research Model

In the adaptive learning system, learning features as a large database system is a huge data resource, every day a large number of records stored in the database, which may include some duplication, irrelevant, or even with each other contradictory records. Education experts also present the characteristics of students vary, causing the learner characteristics of the field of education there is no uniform standard of evaluation. Therefore, a large number of the learning log for statistical analysis to identify the greatest impact on learners, most diagnosis of learning characteristics. This synthesis tool using artificial data generated experimental data, developed an adaptive learning system for association mining algorithm, and the mining algorithm is applied to adaptive systems, adaptive learning system to gradually correct the model of learner characteristics, so that students Better use of the system to learn.Adaptive learning system based on the characteristics of the learner characteristics, the paper FP-Tree algorithm is improved. First, from the algorithm to improve its own, for the problem of too many frequent itemsets, FP-growth proposed on the basis of the key items to improve extraction algorithm KEFP-growth, ignored in the analysis do not care about the frequent itemsets. Then, from the data source were improved, for the data source is too large mining inefficiencies can not even in memory load FP-Tree problem, we propose data projection method, is to use divide and conquer idea, the database frequent 1 - itemsets Divided into various frequent 1 - itemset subset of the database, and then were digging a subset of the database, and then combine them. Finally, KEFP-growth algorithm and the combination of projection algorithm, both to eliminate meaningless mining frequent items, but also a large amount of data when the data can be divided. This article also improved through the experimental comparison of three algorithms and the original FP-Tree algorithm performance, experiments show that the use of KEFP-growth algorithm and database projection method algorithm combines adaptive learning system best suited to the mining characteristics of learners.