CAJ | 학술논문

针对SOM 神经网络算法复杂度高精度低以及K-Means聚类算法需事先确定聚类（簇）数目和随机选取初始聚类中心的不足，论文提出了一种SOM神经网络与K-M eans相结合的S-K二次聚类算法，进行功能互补。该算法应用在SM T焊接质量上，能提高数据聚类信息的精确度，直观地看到数据的分布情况，改善系统的整体性能。
침대SOM 신경망락산법복잡도고정도저이급K-Means취류산법수사선학정취류（족）수목화수궤선취초시취류중심적불족，논문제출료일충SOM신경망락여K-M eans상결합적S-K이차취류산법，진행공능호보。해산법응용재SM T한접질량상，능제고수거취류신식적정학도，직관지간도수거적분포정황，개선계통적정체성능。
As the high complexity and low precision of the SOM neural network algorithm and the shortcomings of the K-Means clustering algorithm needs to determine the number of clusters advanced and randomly select initial clustering cen-ters ,a SOM neural network combined with K-Means of S-K secondary clustering algorithm is proposed ,having complemen-tary function .If the algorithm is used in SMT soldering ,the precision of data clustering information can be enhanced .The intuitive vision of data distributioncould be got and the overall performance of the system is improved .