【问题标题】:Why is my neuralnet function not working in my program?为什么我的神经网络功能在我的程序中不起作用?
【发布时间】:2021-12-30 17:06:49
【问题描述】:

我收到了错误Error in eval(predvars, data, env) : object 'B' not found,我不知道该怎么做:

nn <- neuralnet(B+M~ area+texture+smoothness, data=cancertrain, hidden=3,

B+M是两个潜在的值,要么是良性的,要么是恶性的,对判断影响更大的三个属性是面积、纹理和平滑度。我假设我只是错误地完成了函数神经网络中的参数,有人知道吗?这是公开的 Google 电子表格中的cancer dataset。

library(neuralnet)
library(ISLR) 
library(rpart)
library(rpart.plot)
library(caTools)
library(random)

#setwd("**change to your working directory**")
data <- read.csv("WDBC.csv", header=T)
#head(data)

cancer.dataset <- data
  
#according to previous models and studies, area, texture, and smoothness are the
#attributes with the highest relevance to the diagnosis of benign or malignant
cancer.dataset$b <- cancer.dataset$Diagnosis == "B"
cancer.dataset$m = cancer.dataset$Diagnosis == "M"
cancer.dataset$area <- cancer.dataset$Diagnosis == "area"
cancer.dataset$texture = cancer.dataset$Diagnosis == "texture"
cancer.dataset$smoothness = cancer.dataset$Diagnosis == "smoothness"

cancerdata <- data.frame(cancer.dataset$Diagnosis, cancer.dataset$texture, cancer.dataset$smoothness, cancer.dataset$area)
cancerdata

train <- sample(x = nrow(cancerdata), size = nrow(cancerdata)*0.5)
train

cancertrain <- cancer.dataset[train,]
cancervalid <- cancer.dataset[-train,]
print(nrow(cancertrain))
print(nrow(cancervalid))
nn <- neuralnet(B+M~ area+texture+smoothness, data=cancertrain, hidden=3,  
                rep = 2, err.fct = "ce", linear.output = F, lifesign = "minimal", stepmax = 10000000)

这是教授给出的正确示例使用标准 Iris 数据集的样子,我不确定根据这个示例是否正确:

iris.dataset$setosa <- iris.dataset$Species=="setosa"
iris.dataset$virginica = iris.dataset$Species == "virginica"
iris.dataset$versicolor = iris.dataset$Species == "versicolor"
train <- sample(x = nrow(iris.dataset), size = nrow(iris)*0.5)
train
iristrain <- iris.dataset[train,]
irisvalid <- iris.dataset[-train,]
print(nrow(iristrain))
print(nrow(irisvalid))
nn <- neuralnet(setosa+versicolor+virginica ~ Sepal.Length + Sepal.Width, data=iristrain, hidden=3,  
                rep = 2, err.fct = "ce", linear.output = F, lifesign = "minimal", stepmax = 10000000)

plot(nn, rep="best")

【问题讨论】:

    标签: r deep-learning neural-network iris-dataset


    【解决方案1】:

    您可以使用以下代码

    library(neuralnet)
    library(ISLR) 
    library(caTools)
    library(random)
    
    #setwd("**change to your working directory**")
    data <- read.csv("WDBC.csv", header=T)
    head(data)
    
    #Select the important variables
    cancerdata <- subset(data, select = c(Diagnosis, texture, smoothness, area))
    head(cancerdata) 
    
    train <- sample(x = nrow(cancerdata), size = nrow(cancerdata)*0.5)
    
    cancertrain <- cancerdata[train,]
    cancervalid <- cancerdata[-train,]
    print(nrow(cancertrain))
    print(nrow(cancervalid))
    
    nn <- neuralnet(Diagnosis ~ area+texture+smoothness, data=cancertrain, hidden=3,  
                    rep = 2, err.fct = "ce", linear.output = F, lifesign = "minimal", stepmax = 10000000)
    
    plot(nn, rep="best")
    

    【讨论】:

      猜你喜欢
      • 2019-08-16
      • 2019-04-08
      • 1970-01-01
      • 2018-02-11
      • 1970-01-01
      • 2021-10-15
      • 2013-05-12
      • 2012-07-25
      • 1970-01-01
      相关资源
      最近更新 更多