2017-06-17 80 views
0

當試圖培養我提示以下錯誤:在神經元神經網絡錯誤

錯誤的神經網絡[我]%*%的權重[我]:非順應性參數

nn.model <- neuralnet(formula = Attrition ~ Age + Attrition + BusinessTravel + Department + 
+       DistanceFromHome + Education + EducationField + 
+       EnvironmentSatisfaction + Gender + JobInvolvement + 
+       JobLevel + JobRole + JobSatisfaction + MaritalStatus + 
+       MonthlyIncome + NumCompaniesWorked + OverTime + 
+       PerformanceRating + RelationshipSatisfaction + StockOptionLevel + 
+       TotalWorkingYears + WorkLifeBalance + YearsAtCompany + YearsInCurrentRole + 
+       YearsWithCurrManager , data = TrainData, hidden = 40, 
+       err.fct = "sse", linear.output = FALSE, 
+       lifesign = "full",lifesign.step = 1, 
+       threshold = 0.1) 

樣本數據:

 Attrition   Age BusinessTravel Department DistanceFromHome  Education EducationField 
1781   2 -0.6485577827 -0.5899479808 2.3887402665 -0.88736416029 0.08503478649 2.4710614174 
2852   2 -0.9770074637 2.4160261328 0.4937331474 0.83986126473 0.08503478649 0.3754112538 
513   2 -1.0864906907 -0.5899479808 0.4937331474 -0.76399091564 1.06160616261 0.3754112538 
1398   2 1.8695564380 -0.5899479808 0.4937331474 -0.02375144778 -0.89153658963 -1.0216888553 
2128   2 -0.8675242367 -0.5899479808 0.4937331474 -0.27049793707 -1.86810796575 0.3754112538 
2572   2 -0.5390745557 -0.5899479808 0.4937331474 -0.51724442636 -0.89153658963 -1.0216888553 
    EnvironmentSatisfaction  Gender JobInvolvement  JobLevel  JobRole JobSatisfaction 
1781   -1.5754182775 0.8163577092 -1.0259922033 -0.05777771663 2.3442756382 -1.5676398930 
2852   0.2545816181 0.8163577092 0.3796075533 -0.96132285967 -0.6684127991 0.2461583303 
513   -1.5754182775 0.8163577092 -1.0259922033 -0.96132285967 -0.6684127991 1.1530574420 
1398   -1.5754182775 -1.2245365637 0.3796075533 -0.05777771663 -0.6684127991 0.2461583303 
2128   -1.5754182775 -1.2245365637 -1.0259922033 -0.96132285967 -0.2380287366 1.1530574420 
2572   1.1695815659 -1.2245365637 -1.0259922033 -0.05777771663 -0.6684127991 -0.6607407814 
    MaritalStatus MonthlyIncome NumCompaniesWorked  OverTime PerformanceRating 
1781 0.1332594155 -0.0197425573  0.1228398011 0.6281342689  -0.4261575208 
2852 -1.2366101006 -0.9260223171  -0.6779340687 0.6281342689  -0.4261575208 
513 -1.2366101006 -0.9470541258  -1.036 0.6281342689  -0.4261575208 
1398 0.1332594155 -0.7660531049  0.1228398011 0.6281342689  -0.4261575208 
2128 1.5031289317 -0.8435946221  1.3240006059 0.6281342689  -0.4261575208 
2572 0.1332594155 -0.1327619741  0.1228398011 0.6281342689  -0.4261575208 
    RelationshipSatisfaction StockOptionLevel TotalWorkingYears WorkLifeBalance YearsAtCompany 
1781    1.1912353428 -0.9318558699  -0.2930270824 0.3380386595 -0.817594864371 
2852    0.2661872953 -0.9318558699  -0.8072017302 -1.0776788035 -0.327837549528 
513    1.1912353428 -0.9318558699  -0.8072017302 0.3380386595 -0.491089987809 
1398    0.2661872953  1.4157501777  -0.2930270824 -1.0776788035 -0.491089987809 
2128    0.2661872953  2.5895532015  -0.4215707444 0.3380386595 -0.491089987809 
2572   -1.5839087997  0.2419471539  0.0926039034 0.3380386595 -0.001332672966 
    YearsInCurrentRole YearsWithCurrManager 
1781  -0.6153868953  -1.15573810597 
2852  -0.3393360146  -0.03451387513 
513  -0.6153868953  -0.87543204826 
1398  -0.3393360146  -0.31481993284 
2128  -0.3393360146  -0.31481993284 
2572  -0.8914377760  0.24579218258 

可以請別人出主意,我錯過了什麼。

感謝您期待您的幫助。

問候。 Shaz

回答

0

我的不好。這裏有一個語法錯誤應該是正確的語法:

nn.model <- neuralnet(formula = Attrition ~ Age + Attrition + BusinessTravel + Department + 
+       DistanceFromHome + Education + EducationField + 
+       EnvironmentSatisfaction + Gender + JobInvolvement + 
+       JobLevel + JobRole + JobSatisfaction + MaritalStatus + 
+       MonthlyIncome + NumCompaniesWorked + OverTime + 
+       PerformanceRating + RelationshipSatisfaction + StockOptionLevel + 
+       TotalWorkingYears + WorkLifeBalance + YearsAtCompany + YearsInCurrentRole + 
+       YearsWithCurrManager , data = TrainData, hidden = 40, 
+       err.fct = "sse", linear.output = FALSE, 
+       lifesign = "full",lifesign.step = 1, 
+       threshold = 0.1) 

乾杯。 Shaz