基于改进朴素贝叶斯方法的元器件分类技术  

Component Classification Technology Based on Improved Naive Bayesian Method

作　　者：郑丽香[1] 凌亚东 陈泫文 李颖[1] 刘馨阳 ZHENG Lixiang;LING Yadong;CHEN Xuanwen;LI Ying;LIU Xinyang(CEPREI,Guangzhou 510610,China)

机构地区：[1]工业和信息化部电子第五研究所

出　　处：《电子产品可靠性与环境试验》2020年第1期49-53,共5页Electronic Product Reliability and Environmental Testing

摘　　要：元器件清单数据处理分析是元器件选用管理与状态管控的主要内容。在元器件数据信息化管理的基础上,借助新一代信息技术手段对元器件清单进行处理,可以使得数据处理分析过程更加智能与高效。结合当前人工智能领域的计算机机器学习技术,提出了一种基于改进朴素贝叶斯算法的元器件分类方法,通过加入拉普拉斯平滑方法处理零概率问题、引入对数运算提高计算稳定性等手段,实现了对元器件产品和厂商的自动与智能分类,并对比分析了智能分类结果的准确度,发现运用改进后的算法可提高分类的准确度,对于提高元器件的分类效率,避免人工分类造成的误差具有重要的意义。Component list data processing and analysis is the main content of component selection management and state control.On the basis of the information management of component data,component lists can be processed with the help of the new generation information technology,which can make the data processing and analysis process more intelligent and efficient.Based on the current computer machine learning in the field of artificial intelligence,a component classification method based on improved Naive Bayesian algorithm is proposed.By adding laplacian smoothing method to deal with zero probability problem and introducing logarithmic operation to improve computing stability,the automatic and intelligent classification of component products and manufacturers is realized.The accuracy of intelligent classification results is compared and analyzed,and it is found that the application of improved algorithms can improve the accuracy of classification,which is of great significance to improve the classification efficiency of components and avoid errors caused by manual classification.

关 键 词：元器件 元器件产品分类 元器件厂商分类 朴素贝叶斯 机器学习 

分 类 号：O212.8[理学—概率论与数理统计]