Extract Tidy Fitted Values from an rga Object.

Returns a tidy data frame of fitted cumulative failure counts, confidence bounds, and (when available) Crow-AMSAA model coefficients from an `rga` object. Suitable for exporting results for use in reproducible research workflows.

Usage

tidy_rga(rga_obj)

Arguments

rga_obj: An object of class `'rga'`, or a list of such objects. Each object is created using the [ReliaGrowR::rga()] function.

Value

A named list with two elements:

`fitted`: For a single object, a `data.frame` with columns `time`, `cum_failures`, `fitted`, `lower`, `upper`. For a list, a list of such data frames (one per object).
`params`: For a single object, a one-row `data.frame` with columns `lambda` and `beta` (Crow-AMSAA Power Law parameters), or `NULL` if the model coefficients cannot be extracted. For a list, a list of such data frames.

Details

The Crow-AMSAA (NHPP Power Law) model gives the expected cumulative failures as \(E[N(t)] = \lambda t^\beta\). A \(\beta < 1\) indicates reliability growth (decreasing failure rate); \(\beta > 1\) indicates degradation. The parameters are recovered from the fitted log-log linear model via coef(): \(\lambda = \exp(\text{intercept})\) and \(\beta\) is the slope coefficient.

References

Crow, L. H. (1974). Reliability Analysis for Complex Repairable Systems. In Reliability and Biometry, SIAM, pp. 379-410.

Examples

library(ReliaGrowR)
times <- c(100, 200, 300, 400, 500)
failures <- c(1, 2, 1, 3, 2)
obj <- rga(times, failures)
result <- tidy_rga(obj)
result$fitted
#>   time cum_failures   fitted     lower     upper
#> 1  100            1 1.065602 0.7656742  1.483017
#> 2  300            3 2.562478 2.1157687  3.103503
#> 3  600            4 4.457440 3.7410494  5.311015
#> 4 1000            7 6.703022 5.4093372  8.306102
#> 5 1500            9 9.266388 7.1120671 12.073275
result$params
#>       lambda      beta
#> 1 0.02693046 0.7986756