fuzzySurvival 0.1.0

First release.

Fuzzy numbers

• fuzzy_number_trapezoid(), fuzzy_number_triangle(), fuzzy_number_interval() and fuzzy_number_crisp() build vague quantities.
• alpha_cuts(), support(), core(), membership(), fuzzy_spread(), is_crisp(), as_crisp() and as_fuzzy_number() inspect them.
• Arithmetic + - * / ^ and exp() follow interval arithmetic on the alpha-cuts.
• defuzz() implements nine type-reduction rules.

Data

• fuzzy_surv() bundles fuzzy times with failure degrees, the analogue of survival::Surv().
• sim_fuzzy_surv() generates exact times and then coarsens them the way a record does, for validation studies.
• fuzzy_dfs, a simulated example of 240 subjects with quarterly examinations.

Estimators

• fkm(), the fuzzy Kaplan-Meier estimator, with grouped fits, alpha-cut bounds and a fuzzy survival difference.
• fcox(), the fuzzy Cox proportional hazards model, giving fuzzy coefficients, hazard ratios and p-values.
• flogrank(), a fuzzy log-rank test for two groups.

All three reduce exactly to their survival counterparts on precise data.