广东工业大学学报
廣東工業大學學報
엄동공업대학학보
JOURNAL OF GUANGDONG UNIVERSITY OF TECHNOLOGY
2015年
2期
48-52,103
,共6页
模糊神经网络%沉积环境%判别分析%粒度分析%分类
模糊神經網絡%沉積環境%判彆分析%粒度分析%分類
모호신경망락%침적배경%판별분석%립도분석%분류
fuzzy neural network%sedimentary environment%identification analysis%grain size analysis%classification
由于粒度分析与沉积环境间密切的关系，针对模糊逻辑与人工神经网络各自的优点，提出了一种基于模糊神经网络的沉积环境判别方法．它以碎屑岩的关键粒度参数作为网络的输入，通过标准化和模糊化及输出的去模糊化等过程，使得模糊推理与神经网络充分结合．实验证明，这种模型判别相应沉积环境的误判率为9．1％，明显低于BP神经网络的32．1％且收敛速度更快，更能够满足实际工程的需求．
由于粒度分析與沉積環境間密切的關繫，針對模糊邏輯與人工神經網絡各自的優點，提齣瞭一種基于模糊神經網絡的沉積環境判彆方法．它以碎屑巖的關鍵粒度參數作為網絡的輸入，通過標準化和模糊化及輸齣的去模糊化等過程，使得模糊推理與神經網絡充分結閤．實驗證明，這種模型判彆相應沉積環境的誤判率為9．1％，明顯低于BP神經網絡的32．1％且收斂速度更快，更能夠滿足實際工程的需求．
유우립도분석여침적배경간밀절적관계，침대모호라집여인공신경망락각자적우점，제출료일충기우모호신경망락적침적배경판별방법．타이쇄설암적관건립도삼수작위망락적수입，통과표준화화모호화급수출적거모호화등과정，사득모호추리여신경망락충분결합．실험증명，저충모형판별상응침적배경적오판솔위9．1％，명현저우BP신경망락적32．1％차수렴속도경쾌，경능구만족실제공정적수구．
As to the relationship between grain size and sedimentary environment , in this paper , an iden-tification method of sedimentary environment is proposed based on fuzzy neural network , which combines the advantages of both fuzzy logic and artificial neural network .The proposed approach includes taking the key size parameters of clastic rock as inputs , being standardized and fuzzified by the network and be-ing defuzzified of outputs .As a result , the fuzzy inference process is involved in the neural network suffi-ciently and successfully .The experiment shows that the improved network ’ s misjudgment rate of identifi-cation is 9.1%, less than 32.1%of BPNN obviously.Moreover, the former is faster than the latter in the aspect of convergence .Therefore , the network in this paper can fulfill the necessaries of practical projects.
