Clark

Question 1

Loglogistic

Answer 1

G = x^w/(x^w + o^w)

Question 2

Weibull

Answer 2

G = 1 - e^-(x/o)^w

Question 3

Variance in Clark

Answer 3

o^2 = 1/(n-p) * sum [(actual - expected)^2 / expected]
p = number of parameters (CC = 3, CL = AYs + 2)

Question 4

Loglikelihood in Clark

Answer 4

l = sum (c*ln(expected) - expected)

Question 5

Assumptions in Clark model

Answer 5

Works only for positive expected incremental losses
Variance of incremental losses is proportional to the mean for each cell in the triangle
Incremental losses are iid
Parameter variance is a lower bound based on R-C