The name is absent

89

of As are fairly small, because it represents the “daily” hazard rate of an insurance
company.

The ^uZχ statistic” is 0.984 for λs by coxph method; -8.603 for λ_cs by coxph. We
have λ_o=O.OOOOO8O7, Λ₁=0.0000123; Λ_θc=O.OOO14O and λ^ι_c=0.0000328.

Similarly, the ^tlzχ statistic” is 1.149 for As and -5.196 for λ_cs by the proposed
method. We have λ_o=0.0000673, A_i=O.000103; Aq_c=O.000197 and Aι_c=0.0000489.
Figure 5.1 shows the increase of the monthly hazard rate for bankruptcy based on
the change-point model in Section 5.2.1, i.e. covariates are considered as well. And
Figure 5.2 shows the decrease of the “monthly” hazard rate for acquisition based on
the change-point model in Section 5.2.1.

Test 8 with θ = December 99; τ = -0.5

Table 5.2 shows the parameter estimates for both βs and As.

The “zx statistic” is 0.919 for As by coxph; -8.689 for λ_cs by coxph. We have
Aq=O.00000807, A₁ =0.0000123; A_θc=O.OOO14O and Aι_c=0.0000328. Note that in Table
5.2, the Standard Errors of Coxph estimates are slightly different from the results
when τ=0.5 because the bootstrap method is used to calculate the SE of estimates.

Similarly, the ''^'Zχ statistic” is 0.127 for As and -4.955 for λ_cs by the proposed
method. We have λ^_o=0.0000313, λ^₁=0.0000342; λ_θc=0.000250 and λ^ι_c=0.0000523.
Figure 5.3 shows the increase of the monthly hazard rate for bankruptcy based on
the change-point model in Section 5.2.1, i.e. covariates are considered as well. And

<<   99   100   101   102   103   104   105   >>

More intriguing information