Predict score probabilities from a fitted brs model — brs_predict

Computes predicted probabilities for integer scores on the original scale \(\{0, 1, \ldots, K\}\) implied by the fitted beta interval model.

Usage

brs_predict_scoreprob(
  object,
  newdata = NULL,
  scores = NULL,
  format = c("matrix", "long"),
  id_col = "id"
)

Arguments

object: A fitted "brs" object.
newdata: Optional data frame for prediction.
scores: Optional integer vector of scores to evaluate. Defaults to all scores in 0:ncuts.
format: Output format: "matrix" (default) or "long".
id_col: Column name for observation id when format = "long".

Value

If format = "matrix", a numeric matrix with one row per observation and one column per requested score.

If format = "long", a long data frame with columns id_col, score, and prob.

Details

For a score \(s\) and \(K =\) ncuts, probabilities are computed as:

\(P(Y=s)=F(\mathrm{lim}/K)\) for \(s=0\),
\(P(Y=s)=1-F((K-\mathrm{lim})/K)\) for \(s=K\),
\(P(Y=s)=F((s+\mathrm{lim})/K)-F((s-\mathrm{lim})/K)\) for \(s \in \{1,\ldots,K-1\}\),

where \(F\) is the beta CDF under the fitted \((\mu_i,\phi_i)\).

References

Hawker, G. A., Mian, S., Kendzerska, T., and French, M. (2011). Measures of adult pain: Visual Analog Scale for Pain (VAS Pain), Numeric Rating Scale for Pain (NRS Pain), McGill Pain Questionnaire (MPQ), Short-Form McGill Pain Questionnaire (SF-MPQ), Chronic Pain Grade Scale (CPGS), Short Form-36 Bodily Pain Scale (SF-36 BPS), and Measure of Intermittent and Constant Osteoarthritis Pain (ICOAP). Arthritis Care and Research, 63(S11), S240-S252. doi:10.1002/acr.20543.

Hjermstad, M. J., Fayers, P. M., Haugen, D. F., et al. (2011). Studies comparing Numerical Rating Scales, Verbal Rating Scales, and Visual Analogue Scales for assessment of pain intensity in adults: a systematic literature review. Journal of Pain and Symptom Management, 41(6), 1073-1093. doi:10.1016/j.jpainsymman.2010.08.016.

Examples

# \donttest{
set.seed(33)
dat <- data.frame(x1 = rnorm(100), x2 = rnorm(100))
sim <- brs_sim(
  formula = ~ x1 + x2, data = dat,
  beta = c(0.1, -0.4, 0.3), phi = 0.2, ncuts = 100, repar = 2
)
fit <- brs(y ~ x1 + x2, data = sim, repar = 2)

pmat <- brs_predict_scoreprob(fit)
head(pmat[, 1:5])
#>          score_0     score_1     score_2     score_3     score_4
#> [1,] 0.031430978 0.023248768 0.016145283 0.013227826 0.011526741
#> [2,] 0.074951430 0.040971042 0.026186408 0.020498047 0.017296166
#> [3,] 0.176295577 0.063169280 0.036799106 0.027385744 0.022299473
#> [4,] 0.147774517 0.058628992 0.034844930 0.026215090 0.021510394
#> [5,] 0.004123488 0.004724261 0.003795534 0.003371817 0.003111343
#> [6,] 0.363128384 0.074208336 0.039629538 0.028137509 0.022166138

plong <- brs_predict_scoreprob(fit, scores = 0:10, format = "long")
head(plong)
#>   id score       prob
#> 1  1     0 0.03143098
#> 2  1     1 0.02324877
#> 3  1     2 0.01614528
#> 4  1     3 0.01322783
#> 5  1     4 0.01152674
#> 6  1     5 0.01038186
# }