机器学习-5 线性回归 with R

一元线性拟合

• 数据：见http://pan.baidu.com/s/1qu6dq

• options(scipen = 100)
  
  # 读文件
  city_gdp <- read.table("D:/Sepcial_For_R/city_gdp.txt", 
                         sep = "\t",
                         col.names=c("city", "people", "gdp"), 
                         head = T);
  
  # 散点图，y ~ x
  plot(city_gdp$gdp ~ city_gdp$people)
  
  # 多元线性拟合
  # 线性拟合 lm(formula, data)
  #   formula: 执行拟合的模型形式，y ~ x1+x2+x3
  #   data: 包含拟合模型的数据框
  reg <- lm(city_gdp$gdp ~ city_gdp$people, data = city_gdp)
  
  # 拟合模型的详细结果
  summary(reg)
  
  # 查看拟合模型的预测值
  fitted(reg)
  
  # 查看拟合模型的残差值
  residuals(reg)
  
  # 作图
  plot(city_gdp$gdp ~ city_gdp$people,
       xlab="People",
       ylab="GDP")
  
  # 做拟合的直线
  abline(reg, col='red')