基于熵权灰关联度分析法的煤样识别研究

针对常用煤炭识别方法样本数据要求高、人为因素影响大、识别准确性差等问题,提出一种基于信息熵的灰色关联度分析方法。通过分析熵权灰关联分析法的决策模型,根据样本数据建立了熵权灰关联度分析法的煤炭种类(无烟煤,烟煤和褐煤)辨识模型,并反向检验了待测煤样识别的准确性。最后对比分析了熵权灰关联度分析法和模糊识别法的煤样识别的准确性。结果表明:熵权灰关联度分析法对建模煤样和待测煤样的识别准确率均达到100%;而模糊识别法对建模煤样、待测煤样的识别准确率分别为83.33%、75.00%,总准确率为81.25%。说明熵权灰关联度分析法泛化能力强、可靠性高,明显优于模糊识别法。

Application of entropy coefficient grey correlation analysis in coal sample recognition

The common coal identification methods,which are influenced greatly by human factors,require high level sample data. While the accuracy is poor. Provided a grey correlation analysis method based on the information entropy. Combining the entropy theory and gray relational analysis,established the entropy coefficient grey correlation analysis of coal type identification model which applies to anthracite,bituminous coal and lignite. The accuracy of the model was tested by antidromic method. Taking the modeling coal sample and coal sample under test as research objects,compared the accuracy of entropy coefficient grey correlation degree analysis method and fuzzy identification method. The first analysis method has a high recognition efficiency to the two kinds of coal samples,both the efficiency are 100%. While the efficiency of the second method is 83. 33% and 75. 00%,the total accuracy rate is 81. 25%. The generalization and reliability of the entropy coefficient grey correlation analysis method is high.