Learn R Programming

lifecontingencies (version 1.3.12)

pxt: Functions to evaluate survival, death probabilities and deaths.

Description

These functions evaluate raw survival and death probabilities between age x and x+t

Usage

dxt(object, x, t, decrement)
pxt(object, x, t, fractional = "linear", decrement)
qxt(object, x, t, fractional = "linear", decrement)

Value

A numeric value representing requested probability.

Arguments

object

A lifetable object.

x

Age of life x. (can be a vector for pxt, qxt).

t

Period until which the age shall be evaluated. Default value is 1. (can be a vector for pxt, qxt).

fractional

Assumptions for fractional age. One of "linear", "hyperbolic", "constant force" (can be abbreviated).

decrement

The reason of decrement (only for mdt class objects). Can be either an ordinal number or the name of decrement

Author

Giorgio A. Spedicato

Warning

The function is provided as is, without any warranty regarding the accuracy of calculations. The author disclaims any liability for eventual losses arising from direct or indirect use of this software.

Details

Fractional assumptions are:

  • linear: linear interpolation between consecutive ages, i.e. assume uniform distribution.

  • constant force of mortality : constant force of mortality, also known as exponential interpolation.

  • hyperbolic: Balducci assumption, also known as harmonic interpolation.

Note that fractional="uniform", "exponential", "harmonic" or "Balducci" is also authorized. See references for details.

References

Actuarial Mathematics (Second Edition), 1997, by Bowers, N.L., Gerber, H.U., Hickman, J.C., Jones, D.A. and Nesbitt, C.J.

See Also

exn, lifetable

Examples

Run this code
	#dxt example
	data(soa08Act)
	dxt(object=soa08Act, x=90, t=2)
	#qxt example
	qxt(object=soa08Act, x=90, t=2)
	#pxt example
	pxt(object=soa08Act, x=90, t=2, "constant force" )
	#add another example for MDT

Run the code above in your browser using DataLab