基于深度学习的不确定数据频繁项集挖掘系统

Uncertain data frequent itemset mining system based on deep learning

传统不确定数据频繁挖掘系统工作过程花费的时间较长,且挖掘结果与真实结果误差较大。为了解决上述问题,基于深度学习研究了一种新的不确定数据频繁挖掘系统,在硬件结构中建立深度学习挖掘模型,通过传感器、隐层、输入层、输出层、中心处理器、存储器和显示器构成硬件架构,软件流程由发送采集命令、预训练、微调训练、数据检测、判断候选项集是否为频繁项集等步骤组成。为检测挖掘系统工作性能,与传统挖掘系统进行实验,结果表明,基于深度学习的不确定数据频繁挖掘系统能够在短时间内取得有效的挖掘结果,误差小,实用性更强。

The traditional uncertain data frequent mining system spends less time in working process,and the error between mining results and real results is large. In order to solve the above problems,a new uncertain data frequent mining system is studied based on deep learning. A deep learning mining model is established in the hardware structure. The hardware structure is composed of sensors,hidden layer,input layer,output layer,central processor,memory and display. The software flow is composed of sending acquisition commands,pre-training,fine-tuning training,data detection and judgment.Whether the candidate itemsets are frequent itemsets or not. In order to test the performance of mining system,experiments with traditional mining system show that the uncertain data frequent mining system based on deep learning can obtain effective mining results in a short time,with small error and stronger practicability.