模糊神经网络工具箱应用

《模糊神经网络工具箱应用》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、模糊神经网络工具箱应用一、数据归一化f个样本第j个指标的平均值：标准差为：原始数据标准化为：二、示例数据1.dat训练样本0.58660.06130.62830.11940.79450.639340.59240.04930.330630.28740.04670.719210.0680.78020.495240.2920.24650.85610.060830.01350.549430.79590.51930.295620.1290.88420.545340.11060.99740.650540.52

2、90.09030.40530.02770.95160.637340.16260.9080.560540.77720.9258020.70720.84120.016420.47680.11350.755210.62270.18770.570930.74570.87590.2252210.87620.042820.57050.16110.531230.56630.00970.401530.950.75810.489320.05080.79310.605340.69720.8530.243620.487

3、30.04390.968710.55630.08110.349330.10860.78910.642540.09130.79880.524440.0730.78960.57944注：前三个数据为归一化后的数值，第四个数为所属类别。2.dat测试样本0.57380.0430.517610.63380.15790.485910.28840.40380.67510.61790.08110.36110.14930.89190.552710.80370.92950.148510.768610.138710.

4、60160.02120.582610.05980.87930.644610.50270.26670.680210.53780.18920.537110.53840.18890.367310.41680.13360.980810.79050.56230.359310.11280.90510.620510.72090.49040.018910.13750.8040.513910.507100.478910.1090.89020.509610.74470.85120.135110.7550.78960.

5、257710.5320.10120.650710.66270.01980.438910.50290.2078110.37650.01950.8109100.98170.543210.42580.18370.731710.60790.08770.48310.61310.19230.605310.62790.11720.45221三、建立模糊神经网络过程1.在MATLAB命令行中输入：anfisedit图12.载入数据，选择“1.dat”图23.勾选“Gridpartition”,单击“Generat

6、eFIS”，出现图4，参数如图设置设置好后，单击“ok”，返回。图3图44.执行“Edit”——“FISProperties”，出现图5，如图设置，设置完成单击“Close”。图55.单击“Structure”，观看神经网络结构。图6图76.如图进行参数设置，完成后单击“TrainNow”图8图97.（1）图10为误差曲线，如图11设置，单击“TestNow”，查看训练效果（效果良好），图12。图10图11图12（2）执行菜单命令Edit——MembershipFunctions,查看FIS系统三

7、个输入变量的隶属度函数。（3）执行菜单命令Edit——rules，查看模糊规则8.如图13设置，先单击“ClearPlot”（图14），再单击“Loaddata”，载入“2.dat”（图15）图13图14图159.单击“TestNow”，结果如图17所示图16图17