基于SPSS的贝叶斯逐步线性判别法在煤炭种类识别中的应用

The application of SPSS based Bayes stepwise linear discriminant method in coal species recognition

利用统计软件SPSS的逐步判别分析功能筛选出判别无烟煤、烟煤、褐煤的主要指标——氢含量和氧含量,以该指标为变量建立贝叶斯逐步线性判别函数,并采用该函数分别对建模样本和测试样本进行识别。识别结果显示:基于SPSS软件的贝叶斯逐步线性判别法的分类识别结果正确率为100%,优于模糊识别等算法。SPSS软件具有建模速度快、易操作、精确度高等优点,值得推广和应用。

Using the stepwise discriminant analysis function based on SPSS statistical software to filter out the main indicators of hydrogen and oxygen content of anthracite,bituminous coal and lignite,and based on the index building Bayesian（Bayes） stepwise linear discriminant function,and using the function to recognition modeling and testing samples respectively.The results show that classification correct rate is 100%,better than the fuzzy recognition algorithm.The SPSS software has the advantages of fast in modeling,easy to operation and high accuracy,so it is worth to promote and application.