REPOSITORY CODING · MEDIUM
Respect delayed entry in every risk set
Respect delayed entry in every risk set. Subjects who enter after an event are not at risk for it. The complete application includes native models, services, repository projections, a working preview and visible acceptance cases. Preserve the other documented invariants and run bash test/run.sh.
survival-analysisKaplan Meier survival laboratory
Work through a real codebase.
Choose an environment, investigate the existing implementation and use the tests to check your change. Sign in to open a workspace; execution and Ono usage follow your plan’s allowances.
Task requirements and environments
PySpark · 55 minute session
- Include subjects failing or censoring at a time in that time risk set. Process tied failures before removing censors.
- Respect delayed entry in every risk set. Subjects who enter after an event are not at risk for it.
- Use Greenwood variance S squared times sum d divided by n times n minus d, with terminal extinction variance zero.
- The Kaplan-Meier median is the first time survival is at most one half, and survival updates use the full risk-set fraction.
- Observe events at the inclusive horizon and preserve the censor count in each output point. Future exits become administrative censoring.
- At most forty distinct subjects and nonnegative integer planning times. Cohorts remain separate. Reports round survival and variance to six decimal places; no confidence model is implied beyond the documented estimator. Spark performs risk-set joins and event grouping.