54
For time t, we have
S,(i) = exp{-.5t2exp(∕¾B + ∕¾Age)}, (3.21)
R(t) = exp{-.5texp(∕¾cB + ∕¾cAge)}, (3.22)
where β1 = .5, β2 = .4, βlc = .6 and ∕⅜c = ∙3.
The joint distribution of T and C is specified by the Frank copula as below:
— l)(ɑɑ(ɑ) — 11
•Ж c; ɑ) = logα{l + -----ɪ-----¾ (3.23)
α-l
We tested Kendall’s τ — 0.8, corresponding to a = 0.00000001258 with different
percentages of events, dependent and independent censoring.
In order to generate the observations of event time and informative censoring time,
we used the results by Nelsen (1986):
• Step 1, generate two independent Uniform(0,l) random variables и and v .
. Step 2, let v = logα{l + ,⅛-,1.> tt}.
Then we get the cumulative distribution functions и and v for two variables, which
satisfy the relationship specified by the Frank copula. We get S(f) and R(t) through
и and v. Plug them into Equations 3.21 and 3.22, and then get ⅛ and tc as shown
below:
= /-21og(5(⅛))
τ у exp(Z'iβ')
= -21og(B(t))
c exp(W-βc)
More intriguing information
1. Quality Enhancement for E-Learning Courses: The Role of Student Feedback2. Population ageing, taxation, pensions and health costs, CHERE Working Paper 2007/10
3. The effect of globalisation on industrial districts in Italy: evidence from the footwear sector
4. A Principal Components Approach to Cross-Section Dependence in Panels
5. The name is absent
6. The name is absent
7. The name is absent
8. Commitment devices, opportunity windows, and institution building in Central Asia
9. Campanile Orchestra
10. Imitation in location choice