摘要
为准确地判别矿井水源的类型以减少矿井水害的发生,提出一种改进贝叶斯判别的矿井水源识别模型。通过对砂岩裂隙水、老空水、奥灰水和太灰水4类水源进行水质化验,分析选取K^(+)+Na^(+),Ca^(2+),Mg^(2+),SO_(4)^(2-),Cl^(-),HCO_(3)^(-)6种水质离子作为判别指标;首先使用SPSS Statistics 24软件分析各水质离子之间的相关性,其次对各主成分进行方差贡献率分析,选取前5种水质离子作为主要水质离子,然后根据变异系数法计算主要水质离子权重,最终结合贝叶斯判别法建立水源判别模型,并将模型的预测结果与基础贝叶斯模型的结果进行对比。结果表明:利用改进贝叶斯判别的矿井水源识别模型对14个待测样本进行测试,判别准确率为85.71%,相较于基础贝叶斯模型的准确率提高了21.42%,应用该判别模型的准确率得到了大幅提升;将该模型回代到26个样本中,判别结果与实际情况基本吻合。通过2种模型的对比分析,采用改进贝叶斯模型进行矿井水源识别准确率高且具有研究价值,为矿井水源识别提供新的思路。
In order to reduce the occurrence of mine water disaster,it is necessary to determine the type of mine water source accurately.A mine water source identification model based on improved Bayesian discrimination method is proposed.Six water ions like K^(+)+Na^(+),Ca^(2+),Mg^(2+),SO_(4)^(2-),Cl^(-),HCO_(3)^(-) are selected as the discriminant indicators by testing the water quality of four types of water sources,namely Sandstone fissure water,Goaf water,Ordovician limestone water and Taiyuan formation limestone water.Firstly,the correlation between water ions is analyzed by using SPSS Statistics 24 software.Secondly,the variance contribution rate analysis of each principal component is examined with the top of five water quality ions selected as the main water quality ions.Then,the weight of main water quality ions are determined according to the coefficient of variation.Finally,a water source identification model is established by using Bayesian method and the prediction results of the model are compared with the results of the basic Bayesian model.The results show that 14 samples are tested by the improved Bayesian discrimination model,and the accuracy of the discrimination is 85.71%,which is 21.42%higher than that of basic Bayesian model.The accuracy rate of applying the model has been greatly improved.Twenty-six back-substitution samples are tested with the model,and the discriminant results much the same with the actual situation.The comparison between the two models shows that it is more accurate and valuable to use improved Bayesian model for mine water source identification,which provides a new idea for mine water source identification.
作者
秋兴国
刘杰
李娜
黄润青
QIU Xingguo;LIU Jie;LI Na;HUANG Runqing(College of Computer Science and Technology,Xi’an University of Science and Technology,Xi’an 710054,China)
出处
《西安科技大学学报》
CAS
北大核心
2022年第2期237-244,共8页
Journal of Xi’an University of Science and Technology
基金
国家自然科学基金项目(62002285)。
关键词
主成分分析
变异系数
贝叶斯判别
水源识别
principal component analysis
variation coefficient
Bayesian discrimination
water source identification