R/compute_ret.R

In bczekanski/project7:

compute_ret <- function(y) {
  library(zoo)
  prior_returns <- function(y, months){
    y %>%
      group_by(symbol) %>%
      mutate(returns = roll_mean((monthly_returns), months, fill = NA, align = "right")) %>%
      rename(c(returns = paste0(months, "_month_prior_returns")))
  }

  for(i in c(1, 2, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, 36)) {prior_returns(y, i) -> y}
  y <- y%>%
    gather(key = j.ret, value = prev.returns, 7:ncol(y))
  y$j.ret  <- extract_numeric(y$j.ret)

  prior_returns2 <- function(y, months){
    y %>%
      group_by(symbol) %>%
      mutate(mean_TV = roll_mean((monthly_TV), months, fill = NA, align = "right")) %>%
      rename(c(mean_TV = paste0(months, "_month_prior_volume")))
  }

  for(i in c(1, 2, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, 36)) {prior_returns2(y, i) -> y}
  y <- y%>%
    gather(key = j.vol, value = prev.vol, 9:ncol(y))
  y$j.vol  <- extract_numeric(y$j.vol)

  y <- y %>%
    filter(j.ret == j.vol)


  prior_returns3 <- function(y, months){
    y %>%
      group_by(symbol) %>%
      mutate(returns2 = roll_mean(lead(monthly_returns, n = 1), months, fill = NA, align = "left")) %>%
      rename(c(returns2 = paste0(months, "_month_future_returns")))
  }

  for(i in c(1, 2, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, 36)) {prior_returns3(y, i) -> y}
  y <- y %>%
    gather(key = k.ret, value = future.returns, 11:ncol(y))
  y$k.ret  <- extract_numeric(y$k.ret)

prior_returns4 <- function(y, months){
    y %>%
      group_by(symbol) %>%
      arrange(monthYear) %>%
      mutate(vol2 = roll_mean(monthly_TV,  months, fill = NA, align = "left")) %>%
      rename(c(vol2 = paste0(months, "_month_future_volume")))
  }

  for(i in c(1, 2, 3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, 36)) {prior_returns4(y, i) -> y}
  y <- y %>%
    gather(key = k.vol, value = future.vol, 13:ncol(y))
  y$k.vol  <- extract_numeric(y$k.vol)


  # There needs to be some sort of lag
  # 3-1 the j=3 is from 1-1 to 3-31, while k = 3 is 3-1 to 5-30

  z <- y %>%
    mutate(month = month(monthYear)) %>%
    filter(month == 1) %>%
    #group_by(symbol) %>%
    #filter(monthly_returns <= 200) %>%
    na.omit

}