人工智能知識及應用證書（兼讀制）實習三

Case: Diabetic Perdition in R

Please complete the following task exercises.
Ref: Page232

Task 1: Data Retrieval

Complete the following code

url='https://raw.githubusercontent.com/MicrosoftLearning/mslearn-dp100/main/data/diabetes.csv'
df=read.csv(url ,header=????)

Task 2: Data Checking

Check the data
Number of observations _____
Write down a command to show part of data:
______________________________

Task 3: Normalization

What is normalizaton: See Lecture Note 243
Complete the following code
X <- (X - min(X)) / (max(X) - min(X)) Y <- (Y - min(Y)) / (max(Y) - min(Y))

#df.n = df
#df.n[ , 2] = (df[ , 2]- min(df[ , 2])) / (max(df[ , 2])-min(df[ , 2]))
#???
#???
head(df.n)

Task 4: Build Neuralnet Formula

Complete the following code

print(colnames(df.n))

 [1] "PatientID"              "Pregnancies"           
 [3] "PlasmaGlucose"          "DiastolicBloodPressure"
 [5] "TricepsThickness"       "SerumInsulin"          
 [7] "BMI"                    "DiabetesPedigree"      
 [9] "Age"                    "Diabetic"              

f='Diabetic~????'

Task 5: Prepare Training and Testing Data

Execute the following code

rows  = sample(nrow(df.n))
df.ns = df.n[rows, ]
df.ns.train=df.ns[1   :700 , ]
df.ns.test=df.ns[700:1000 , ]

Task 6: Build/Plot the Model

Execute the following code

#model = neuralnet::neuralnet(f, data = df.ns.train, 
#                           hidden = c(?,?), linear.output = FALSE)

plot(model)

Task 7: Find the Accuracy

data.frame(predict=round(predict(model, df.ns.test),0) , actual=df.ns.test$Diabetic)

#total_testing_obs = ???
#accuracy = sum(round(predict(model, df.ns.test),0) == df.ns.test$Diabetic)/ total_testing_obs
print(accuracy)

[1] 0.8533333