基于大数据的船舶交通客流特征组合预测模型分析
|
摘要
为了提高船舶交通客流特征预测的时效性,设计基于大数据相关技术信息,提出将粗糙集和支持向量机预测机制结合的预测分析模型。首先运用粗糙集属性,对大数据下的船舶交通客流信息,进行数据出行约简,删除数据中冗余属性,继而建立支持向量机回归预测机制,将约简后的船舶交通数据样本,作为数据预处理器,通过对条件值进行筛选,并量化为一张二维表格,作为决策表,重新组合成为训练数据样本,输入SVM中,进行学习训练,实现交通客流特征的组合预测。仿真实验表明,该模型预测结果特征比真实性提高29%,有效时序性提高35%,可以证明该预测模型的预测结果时效性更强。
in order to improve the timeliness of prediction of ship traffic passenger flow characteristics, a prediction analysis model combining rough set and support vector machine(SVM) prediction mechanism was proposed based on big data related technical information., first of all, using rough set attribute for large vessel traffic passenger flow information data, data reduction, travel by reduction algorithm, delete the redundant attributes in the data, and then set up the mechanism of support vector machine(SVM) regression prediction, after reduction of vessel traffic data samples, as the data preprocessor, filtered through the condition value, and quantify as a two-dimensional table, as a decision table, reassembled as training data sample, input of SVM, learning training, to realize traffic flow characteristics of combination forecast. Simulation results show that the characteristics of the predicted results of the model are 29% higher than the authenticity, and the effective timing is 35% higher, which can prove that the predicted results of the prediction model are more time-efficient.
in order to improve the timeliness of prediction of ship traffic passenger flow characteristics, a prediction analysis model combining rough set and support vector machine(SVM) prediction mechanism was proposed based on big data related technical information., first of all, using rough set attribute for large vessel traffic passenger flow information data, data reduction, travel by reduction algorithm, delete the redundant attributes in the data, and then set up the mechanism of support vector machine(SVM) regression prediction, after reduction of vessel traffic data samples, as the data preprocessor, filtered through the condition value, and quantify as a two-dimensional table, as a decision table, reassembled as training data sample, input of SVM, learning training, to realize traffic flow characteristics of combination forecast. Simulation results show that the characteristics of the predicted results of the model are 29% higher than the authenticity, and the effective timing is 35% higher, which can prove that the predicted results of the prediction model are more time-efficient.
引文
[1]张丹丹,王雷.船舶交通特征统计分析中的大数据挖掘应用[J].舰船科学技术, 2017, 39(4A):127–129.
[2]陈欢,薛美根.大数据环境下上海市综合交通特征分析[J].城市交通, 2016, 14(1):24–29.
[3]郎振红.基于云计算自主学习平台的设计[J].电子设计工程,2016, 6(1):35–39.
[4]张红斌.物联网络中对入侵信号优化预警仿真研究[J].计算机仿真, 2016, 33(11):333–336.
[2]陈欢,薛美根.大数据环境下上海市综合交通特征分析[J].城市交通, 2016, 14(1):24–29.
[3]郎振红.基于云计算自主学习平台的设计[J].电子设计工程,2016, 6(1):35–39.
[4]张红斌.物联网络中对入侵信号优化预警仿真研究[J].计算机仿真, 2016, 33(11):333–336.