Stock Linear Regression

学术

R codes for stock estimation regression

• 导入数据及必要模块

    load("F:/Research Issues/Rgis/Countries/stock.RData")
    library(ggplot2)
    library(xtable)

• 数据清洗

    ## data filtering
    al_global <- subset(al_global, light > 0 & stock > 0)
    cement_global <- subset(cement_global, light > 0 & stock > 0)
    steel_global <- subset(steel_global, light > 0 & stock > 0)
    ## select a year for test
    i <- 1992
    al_year <- subset(al_global, year == i)
    cement_year <- subset(cement_global, year == i)
    steel_year <- subset(steel_global, year == i)

• 建立训练集和测试集

    ## divide data t make train set and test set
    set <- 1:length(al_year[,1])
    train_sample <- sample(set, 100)
    al_train <- al_year[train_sample,]
    al_test <- al_year[-train_sample,]

• 对训练集进行线性回归

    ## generate double-log linear model
    al_lm <- lm(log(al_train$stock) ~ log(al_train$light))
    al_coef <- coef(al_lm)