library(MASS)
data1<-Boston
cr<-cor(data1)
corrplot::corrplot(cr,type = 'lower')
corrplot::corrplot(cr,method = 'number')
split<-sample(2,nrow(data1),replace = T,prob=c(0.8,0.2))
train<-data1[split==1,]
test<-data1[split==2,]
model<-lm(medv~.-indus-age,data = train)
summary(model)
##
## Call:
## lm(formula = medv ~ . - indus - age, data = train)
##
## Residuals:
## Min 1Q Median 3Q Max
## -15.2504 -2.8666 -0.5864 1.9972 25.9778
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 38.271517 5.798273 6.601 1.30e-10 ***
## crim -0.105737 0.034812 -3.037 0.00254 **
## zn 0.042153 0.014996 2.811 0.00518 **
## chas 2.289875 1.063797 2.153 0.03195 *
## nox -18.612367 4.103161 -4.536 7.58e-06 ***
## rm 3.792754 0.454692 8.341 1.19e-15 ***
## dis -1.558189 0.214532 -7.263 1.98e-12 ***
## rad 0.295983 0.071706 4.128 4.46e-05 ***
## tax -0.011133 0.003754 -2.966 0.00320 **
## ptratio -0.995191 0.146563 -6.790 4.04e-11 ***
## black 0.008952 0.003097 2.891 0.00405 **
## lstat -0.524153 0.053588 -9.781 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 4.912 on 402 degrees of freedom
## Multiple R-squared: 0.7341, Adjusted R-squared: 0.7268
## F-statistic: 100.9 on 11 and 402 DF, p-value: < 2.2e-16
prdeic<-predict(model,test)
plot(test$medv,type ="l",lty=1.8,col="green")
lines(prdeic,type="l",col="blue")
