TFDEA (version 0.9.8.3)

TFDEA: Technology Forecasting Using DEA

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

rl{ vrs Variable returns to scale, convexity and free disposability drs Decreasing returns to scale, convexity, down-scaling 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 * l

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 front

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)