基于云边协同的决策树并行化设计

Parallel design of decision tree based on cloud side collaboration

随着海量数据的分析任务越来越重,数据挖掘工作需要进一步推进和优化。文章首先提出了基于云边协同的决策树并行化设计,根据连续属性离散化判断分裂属性,在属性确认之后建立决策树;其次对并行化设计内的数据进行预处理,构建决策树整体并行流程;最终实现数据的实时分析与智能处理。对比试验表明,基于云边协同的决策树算法连续属性离散化的优化,在保证准确率的基础上,能有效地缩短运算时间,提高算法的运算速度。

With the increasingly heavy task of analyzing massive data,data mining needs to be further promoted and optimized.Therefore,this paper firstly proposes a parallel design of decision tree based on cloud edge collaboration;secondly,according to the discretization of continuous attributes,the split attributes are judged,and the decision tree is established after the attributes are confirmed.Preprocess the data in the parallel design,construct the overall parallel process of the decision tree,and finally realize the real-time analysis and intelligent processing of the data.Comparative experiments show that the optimization of continuous attribute discretization of decision tree algorithm based on cloud edge collaboration can effectively shorten the operation time and improve the operation speed of the algorithm on the basis of ensuring the accuracy.