TFDEA: Technology Forecasting Using DEA

In TFDEA: Technology Forecasting using DEA (Data Envelopment Analysis)

Description

Calculate the technology rate of change (ROC) which can then be used for predicting future product performance, estimate new product release dates, and other purposes.

Usage

TFDEA(x, y, dmu_date_rel, date_forecast, rts="vrs", orientation="output",
  second="min", mode="static", segroc=FALSE, debug=1)

Arguments

x	Inputs or resources used by each decision making unit
y	Outputs or products of each decision making unit
dmu_date_rel	Date of introduction (release) for decision making unit or product
date_forecast	Date of forecast
rts	Returns to scale for the application, production technology, or industry studied vrs Variable returns to scale, convexity and free disposability drs Decreasing returns to scale, convexity, down-scaling and free disposability crs Constant returns to scale, convexity and free disposability irs Increasing returns to scale, (up-scaling, but not down-scaling), convexity and free disposability
orientation	Orientation of the DEA model - primary emphasis on input-reduction input or output-augmentation output. Note that unlike the DEA functions, the default is output orientation.
second	Optional: Enables an alternate secondary objective function based on the product of lambda and dmu_date_rel. The default is min. Other options include none or max which will then skip this step or minimize date * lambda while holding the efficiency constant.
mode	Declares if the technology forecast is done using static or dynamic frontiers. A static frontier is where the forecast is made using a fixed date of date_forecast. A dynamic frontier allows for different based dates on the frontier based on the product of dmu_date and lambda. When a dynamic frontier is selected, a secondary objective function should be specified to avoid problems with multiple optima (second=min or max)
segroc	Uses segmented rate of change if TURE.
debug	Optional: Only for debugging. If debug is a integer greater then zero debug information will be output.

Value

TFDEA returns sets of values at three points in time, at the initial release of the product, _rel, at the forecast data specified by date_forecast, _cur, and for the date in the future when the product is forecast for.

The function returns a number of values per product (DMU). The standardized efficiency (all inefficiencies are between 0 and 1, for input and output orientation) eff, and the lambda values, lambda, are returned.

A rate of technology change roc is returned for products efficient at release. At current time a local rate of change and at forecast time an individualized rate of change is returned - sroc. If segroc = FALSE then the sroc is the average rate of change and is the same for every product.

Lastly a date for current and forecast is returned, date. If mode = dynamic then the current date is the current date adjusted by what products the product is being compared to. If static is used then the date is the date_forecast for all products. If the product release date is in the future then a forecast for the product is returned.

Not all values are calculated for all products at all points in time. For example a a product released in the past with that is inefficient at release would not have a roc or sroc calculated because the product is not used to calculate the overall rate of technology change.

$date_soa	List of unique release dates for conducting analyses
$dmu_eff_rel	Efficiency per product (DMU) at time of release
$dmu_lambda_rel	Lambdas per DMU at time of release
$dmu_eff_cur	Efficiency per product (DMU) at current time (date of forecast)
$dmu_lambda_cur	Lambdas per DMU at current time
$dmu_date_cur	Adjusted current date for DMU
$dmu_roc_cur	Rate of Change for product from date of release to current date
$dmu_sroc_cur	Local rate of change for the product
$dmu_eff_for	Superefficiency per product (DMU)
$dmu_lambda_for	Lambda per DMU at forecast time
$dmu_date_for	Date forecast for product based upon supereffeciency and technology rate of change
$dmu_sroc_for	Individualized rate of change for product
$roc	Average Rate of Technology Change

See Also

DEA

SDEA

Examples

  # Example from Inman (2004) p. 93-104, predicting flash drive introduction dates

  drive <- c("A", "B", "C", "D", "E", "F", "G")

  x           <- data.frame(c(16, 14, 8, 25, 40, 30, 40))
  rownames(x) <- drive
  colnames(x) <- c("Cost")

  y           <- data.frame(c(16, 32, 32, 128, 32, 64, 256))
  rownames(y) <- drive
  colnames(y) <- c("Capacity")


  z           <- data.frame(c(2001, 2002, 2003, 2004, 2001, 2002, 2004))
  rownames(z) <- drive
  colnames(z) <- c("Date_Intro")

  # Calc intro date for products using forecast year 2003
  results <- TFDEA(x, y, z, 2003, rts="vrs", orientation="output", mode="dynamic")

  # Examine what dates are forecast for DMU D & G
  print(results$dmu_date_for)

TFDEA documentation built on May 29, 2017, 11:55 a.m.