摘要
内蒙古地区是沙尘暴频发区,如何利用气象数据通过数据挖掘技术建立高效的沙尘暴预测模型成为了研究者们的热点问题。将具有相对稳定分类效果的朴素贝叶斯分类算法与属性重要度和Adaboost算法框架相结合提出了一种加权的Adaboost-NBC分类方法。上述算法从属性的独立条件与分类决策两方面优化了传统的朴素贝叶斯算法。将改进的朴素贝叶斯算法用于内蒙古地区的沙尘暴预测中。实验结果表明,上述算法相比传统的朴素贝叶斯算法在预测沙尘暴的准确率有所提高。实验结果证明改进的算法确实提高了朴素贝叶斯算法的分类效果及可伸缩性。
Inner Mongolia is a frequent area of sandstorm.How to use weather data to establish an efficient sandstorm prediction model through data mining technology has become a hot issue for researchers.In this paper,we proposed a weighted Adaboost-NBC classification method,which combines the Naive Bayesian classification algorithm with relative stable classification effect,attribute importance and Adaboost algorithm framework.Using this algorithm,we optimized the traditional Naive Bayesian algorithm from two aspects:the independent condition of the attribute and the classification decision.In this paper,the improved Naive Bayesian algorithm was used in the prediction of sandstorms in Inner Mongolia.Through the experimental results,the improved Naive Bayesian algorithm improved the accuracy of sandstorm 11%.The experimental results show that the improved algorithm does improve the classification effect and scalability of the Naive Bayesian algorithm.
作者
仁庆道尔吉
郑碧莹
赵学哲
李娜
REN Qing-dao-er-ji;ZHENG Bi-ying;ZHAO Xue-zhe;LI Na(College of Information Engineering,Inner Mongol University of Technology,Hohhot Inner Mongolia 010051,China)
出处
《计算机仿真》
北大核心
2020年第11期417-421,共5页
Computer Simulation
基金
内蒙古自然科学基金项目(2018MS06021)。
关键词
气象数据
沙尘暴
朴素贝叶斯算法
Meteorological data
Sandstorm
Naive Bayesian classification
