13_AS_5B_lec_annotated

# This may not be possible in practice a simpler

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ˜˜ ˜ Var (λ − µ) ≈ Var (λ) + Var (˜) µ 0.094 0.1 = + 2000 + 3000 + 5000 4000 + 5000 + 3000 = 9.4 × 10−6 + 8.333 × 10−6 = 1.7733 × 10−5 . ˆ and λ and µ are asymptotically normal. Hence, ˆ µ − λ ∼ N 0, 1.7733 × 10−5 55/60 ˆˆ (µ−λ)−0 1−0.094 As Z = √1.7733×10−5 = √10..7733×10−5 = 1.4248 < 1.645. Therefore there is no statistical evidence that λ < µ. Actuarial Statistics – Module 5: Parametric methods: Markov Model General Markov Model Estimation of transition intensities Time in each state for each individual over a study period Can determine the number of jumps out of each state Total waiting time in each state Hence can estimate the transition intensities Then can determine survival function for individuals in each state 56/60 Actuarial Statistics – Module 5: Parametric methods: Markov Model General Markov Model Data issue The calculation of the estimates need to compute the total waiting time. This may not be possible in practice. A simpler approach-the census approac...
View Full Document

{[ snackBarMessage ]}

Ask a homework question - tutors are online