Optimization of FP-Growth algorithm based on cloud computing and computer big data

作     者：Zhang, Baohua 

作者机构：Changzhou Vocat Inst Engn Gen Educ Teaching Dept Changzhou 213164 Jiangsu Peoples R China 

出 版 物：《INTERNATIONAL JOURNAL OF SYSTEM ASSURANCE ENGINEERING AND MANAGEMENT》 (国际系统保证工程与管理杂志)

年 卷 期：2021年第12卷第4期

页      面：853-863页

核心收录：

学科分类：0202[经济学-应用经济学] 02[经济学] 1202[管理学-工商管理] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 0837[工学-安全科学与工程] 0823[工学-交通运输工程] 

基　　金：Institute of higher vocational education, Changzhou University [CDGZ2019043] Analysis and Countermeasure Research on the Application of Online Courses in Internet Plus Age [2020JDKT076] 

主　　题：Cloud computing EI Computer big data FP-Growth algorithm Big data clustering algorithm Optimized design 

摘      要：The rapid development of cloud computing technology has spawned many excellent cloud computing platforms. These cloud computing platforms provide an effective solution for the processing of big data, which can be used as the basis for the study of parallel mining algorithms and the application of algorithms. This article uses the FP-Growth algorithm to mine and analyze computer big data. Aiming at the low extraction efficiency of traditional FP-Growth algorithm in large-scale data environment, an improved FP-Growth algorithm is proposed. In addition, in view of the shortcomings of frequent lists of L elements that are often cross-referenced in the FP-tree construction process, an improved algorithm based on hash tables is proposed, which realizes the storage address processing element name key, and then realizes the element name key to storage numbered mapping. This article mainly introduces the optimization of FP-Growth algorithm under the background of cloud computing and computer big data. The experimental results in this paper show that the performance of the improved FP-gtowth algorithm is better than the original algorithm, the traversal time is reduced by 13%, and the mining efficiency is increased by 25%. In addition, the use of this algorithm for data clustering reduces the error rate and optimizes performance becomes better and has better application value.