rsq_by_lag: Predicts important parameter by the lags
In bookMatrix: Model to build a booking matrix and analyze it

Description Usage Arguments Details Value Examples

View source: R/bookMatrix.R

Fitting important parameter on the target variable by lag

1
2
3

rsq_by_lag(hotelData, r.squared=TRUE, year1=2016, year2=2015, oneyear=FALSE,
                      wdayyes=FALSE, monthyes=FALSE, bayernholiday=FALSE, publicholiday=FALSE,
                      italholiday=FALSE, season=FALSE, season2=FALSE, ...)

`hotelData`	a data.frame with the variables target (contains bookings on arrival date), a undefined number of lags (contains bookings at the days before arrival) and undefined number of control variables
`r.squared`	If `TRUE` r.squared is stored - if `FALSE` lag coefficient is stored
`year1`	The year regressed on
`year2`	Year that is used for the comparison for the same date years before
`oneyear`	If `TRUE` add normalized one year ago on the same date data
`wdayyes`	If `TRUE` add weekdays in regression
`monthyes`	If `TRUE` add month in regression
`bayernholiday`	If `TRUE` add holidays in Bayern in regression
`publicholiday`	If `TRUE` add public holidays in Germany in regression
`italholiday`	If `TRUE` add holidays in Italy in regression
`season`	If `TRUE` add regular season in regression
`season2`	If `TRUE` add shorter season in regression
`...`	arguments to be used to form the default `control` argument if it is not supplied directly.

rsq_by_lag finds parameter for all lags provided

An object of class "data.frame".

##  create sample
date_from <- sort(as.Date(runif(20000, min=1, max=700),origin = "2015-01-01"))
date_creation <- round(runif(20000, min = 1, max = 365), digits = 0) 
date_to <- round(rnorm(20000, mean = 2, sd = 1), digits = 0)    
hotelData2 <- data.frame(date_creation, date_from, date_to)
hotelData2$date_to <- as.Date(hotelData2$date_from + hotelData2$date_to)
hotelData2$date_creation <- as.Date(hotelData2$date_from - hotelData2$date_creation)
hotelData2 <- data.frame(subset(hotelData2, hotelData2$date_to >= hotelData2$date_from & hotelData2$date_from >= hotelData2$date_creation))
rownames(hotelData2) <- seq(length(hotelData2$date_creation),from = 1)
rm(date_creation, date_from, date_to)
hotelDataApply <- lapply(1:nrow(hotelData2), function(i) seq(hotelData2$date_from[i], hotelData2$date_to[i], by = "day"))
hotelData2 <- data.frame(origin = rep(hotelData2$date_creation, sapply(hotelDataApply, length)), target = do.call("c", hotelDataApply))
hotelData2 <- hotelData2[hotelData2$target <= max(hotelData2$origin), , drop = FALSE]
hotelData2$horizon <- as.numeric(hotelData2$target - hotelData2$origin)
cumBookMat2 <- tapply(factor(hotelData2$horizon, levels = 0:max(hotelData2$horizon)), hotelData2$target,
                  function(y) rev(cumsum(rev(table(y)))))
                  cumBookMat2 <- do.call("rbind", cumBookMat2) 
cumBookMat2 <- as.data.frame(cumBookMat2)
names(cumBookMat2) <- c("target", paste0("lag", names(cumBookMat2)[-1]))


## use function with sample hotelData2
rsq <- rsq_by_lag(hotelData=cumBookMat2, r.squared=TRUE)