glmnet錯誤 - 不符合參數

Question

我一直遇到頑固的錯誤，試圖使用glmnet stats package如下所示。

我試過列出的有限建議here（包括將數據設置爲data.matrix）。我也試圖使用？glmnet中描述的「penalty.box」設置，但沒有任何積極的結果。

df = structure(list(term = c(0, 1, 0, 0, 1, 0, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0), int_rate = c(10.65, 15.27, 15.96, 13.49, 12.69, 7.9, 15.96, 18.64, 21.28, 12.69, 14.65, 12.69, 13.49, 9.91, 10.65, 16.29, 15.27, 6.03, 11.71, 6.03, 15.27, 12.42, 11.71, 11.71, 11.71, 9.91, 16.77, 11.71, 11.71, 7.51, 7.9, 15.96, 8.9, 15.96, 10.65, 9.91, 7.9, 12.42, 12.69, 7.51, 7.9, 18.25, 16.77, 6.03, 9.91, 8.9, 10.65, 6.03, 6.62, 9.91), emp_length = c(NA, 1, NA, NA, 1, 3, 8, 9, 4, 1, 5, NA, 1, 3, 3, 1, 4, NA, 1, 6, 3, NA, NA, 5, 1, 2, 2, NA, 1, 7, 5, 2, 2, 7, NA, 2, 1, 1, 1, 4, NA, 9, NA, NA, 6, NA, 6, NA, 5, 8), annual_inc = c(24000, 30000, 12252, 49200, 80000, 36000, 47004, 48000, 40000, 15000, 72000, 75000, 30000, 15000, 1e+05, 28000, 42000, 110000, 84000, 77385.19, 43370, 105000, 50000, 50000, 76000, 92000, 50004, 106000, 25000, 17108, 75000, 29120, 24044, 34000, 41000, 55596, 45000, 36852, 27000, 68004, 62300, 65000, 55000, 45600, 0000, 1e+05, 27000, 60000, 70000, 80000), delinq_2yrs = c(0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 2L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 3L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L), inq_last_6mths = c(1L, 5L, 2L, 1L, 0L, 3L, 1L, 2L, 2L, 0L, 2L, 0L, 1L, 2L, 2L, 1L, 2L, 0L, 0L, 0L, 3L, 0L, 0L, 0L, 1L, 0L, 0L, 0L, 0L, 0L, 0L, 1L, 0L, 1L, 2L, 0L, 0L, 1L, 3L, 0L, 0L, 0L, 0L, 1L, 1L, 0L, 0L, 1L, 0L, 2L), outcome = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)), .Names = c("term", "int_rate", "emp_length", "annual_inc", "delinq_2yrs", "inq_last_6mths", "outcome"), row.names = c(NA, 50L), class = "data.frame") 
X = select(df, -outcome) 
Y = df$outcome 
X_train = as.matrix(X[1:50,]) 
Y_train = as.matrix(Y[1:50]) 

library(glmnet) 

model = glmnet(X_train, Y_train, family = "binomial") 

summary(model) 

這裏的錯誤：

Error in drop(y %*% rep(1, nc)) : 
    error in evaluating the argument 'x' in selecting a method for  
function 'drop': Error in y %*% rep(1, nc) : non-conformable arguments 

實際數據集110個變量和〜1毫米觀察，但部分數據集以上是生產同樣的問題。

對此調試方法的任何建議？

Answer 1

問題是您有NA條目是emp_length列，您不能在您傳遞給glmnet的數據集中包含該列。您或者需要從數據集中刪除包含NA值的行，或者將其忽略。

以下是一些代碼，您可以使用這些代碼刪除值並使glmnet正常工作。

X = select(df, -outcome) 
Y = df$outcome 
X_train = as.matrix(X[1:50,]) 
Y_train = as.matrix(sample(0:1,50,replace=TRUE)) 

has_NA = apply(is.na(X_train), 1, any) #= 1 if any column in that row is NA 
X_train = X_train[!has_NA,] 
Y_train = Y_train[!has_NA,] 

library(glmnet)  
model = glmnet(X_train, Y_train, family = "binomial")