Package {OneShotEM}

Type:	Package
Title:	Efficient eM-Algorithm for One-Shot Device Data Analysis
Version:	0.1.0
Description:	Implements the simple and efficient Expectation-Maximization (eM) algorithm proposed by Zhu, Li, Li, and Balakrishnan (2026) <doi:10.1080/03610918.2025.2515193> for parameter estimation in one-shot device accelerated life testing (ALT) data. Unlike traditional EM algorithms that impute exact failure times, this method treats failure counts between inspection intervals as missing data, resulting in faster convergence and enhanced numerical stability. Supports Exponential, Weibull, Lognormal, Gamma, and custom user-defined lifetime distributions under log-linear stress models. Standard errors, confidence intervals, model selection statistics (AIC, BIC, AICc, HQIC), residual diagnostics, and visualization tools are provided. References: Balakrishnan and Ling (2012) <doi:10.1016/j.csda.2011.09.010>, Fan, Balakrishnan, and Chang (2009) <doi:10.1080/00949650802142592>.
License:	GPL (≥ 3)
LazyData:	true
Depends:	R (≥ 4.0.0)
Imports:	stats, graphics, grDevices, utils, methods, numDeriv
Suggests:	testthat (≥ 3.0.0), knitr, rmarkdown
Config/testthat/edition:	3
Encoding:	UTF-8
Language:	en-US
RoxygenNote:	7.3.3
VignetteBuilder:	knitr
NeedsCompilation:	no
Packaged:	2026-08-07 23:14:09 UTC; shikhar tyagi
Author:	Shikhar Tyagi [aut, cre], Arvind Pandey [aut], Bhupendra Singh [aut], Vrijesh Tripathi [aut]
Maintainer:	Shikhar Tyagi <shikhar1093tyagi@gmail.com>
Repository:	CRAN
Date/Publication:	2026-08-21 12:41:11 UTC

OneShotEM: Efficient eM-Algorithm for One-Shot Device Data Analysis

Description

Implements the simple and efficient Expectation-Maximization (eM) algorithm proposed by Zhu, Li, Li, and Balakrishnan (2026) https://doi.org/10.1080/03610918.2025.2515193 for maximum likelihood parameter estimation in one-shot device accelerated life testing (ALT) data.

Details

One-shot devices (such as electro-explosive devices, automobile airbags, and missiles) can be tested only once, yielding binary data (failed or survived) at pre-determined inspection times. Rather than treating exact failure times as missing data (the traditional EM approach), the eM algorithm treats the failure counts between successive inspection intervals as missing data. This formulation leads to much faster convergence, higher success rates of finding maximum likelihood estimates, and improved numerical stability.

Supported distributions include:

• "exponential": Log-linear scale model

• "weibull": Log-linear scale model with constant or covariate-dependent shape

• "lognormal": Log-linear location model with constant scale

• "gamma": Log-linear scale model with constant shape

• Custom user-defined survival functions

Key functions:

• oneshot_em: Newly proposed eM-algorithm (NPM)

• oneshot_tm: Traditional EM-algorithm (TM)

• oneshot_fit: High-level wrapper function

• em_control: Control parameters for EM algorithms

Author(s)

Maintainer: Shikhar Tyagi shikhar1093tyagi@gmail.com (ORCID)

Authors:

• Arvind Pandey arvindmzu@gmail.com

• Bhupendra Singh bhupendra.rana@gmail.com

• Vrijesh Tripathi vrijesh.tripathi@uwi.edu

References

Zhu, X., Li, Y., Li, T., & Balakrishnan, N. (2026). A simple and efficient eM-algorithm for one-shot device data analysis. Communications in Statistics - Simulation and Computation, 55(3), 959–970. doi:10.1080/03610918.2025.2515193

Balakrishnan, N., & Ling, M. (2012). EM algorithm for one-shot device testing under the exponential distribution. Computational Statistics & Data Analysis, 56(3), 502–509. doi:10.1016/j.csda.2011.09.010

Fan, T., Balakrishnan, N., & Chang, C. (2009). The Bayesian approach for highly reliable electro-explosive devices using one-shot device testing. Journal of Statistical Computation and Simulation, 79(9), 1143–1154. doi:10.1080/00949650802142592

Extract Akaike Information Criterion (AIC)

Description

Extract Akaike Information Criterion (AIC)

Usage

## S3 method for class 'oneshot_em'
AIC(object, ..., k = 2)

Arguments

Value

A numeric scalar of class "numeric" containing the calculated Akaike Information Criterion (AIC) value for the fitted model.

Extract Bayesian Information Criterion (BIC)

Description

Extract Bayesian Information Criterion (BIC)

Usage

## S3 method for class 'oneshot_em'
BIC(object, ...)

Arguments

Value

A numeric scalar of class "numeric" containing the calculated Bayesian Information Criterion (BIC) value for the fitted model.

Extract Model Coefficients

Description

Extract Model Coefficients

Usage

## S3 method for class 'oneshot_em'
coef(object, ...)

Arguments

Value

A named numeric vector of class "numeric" containing the maximum likelihood estimates of the model parameters.

Electro-Explosive Device Accelerated Life Test Data

Description

One-shot device accelerated life test (ALT) dataset reported by Fan, Balakrishnan, and Chang (2009) and analyzed by Zhu et al. (2026).

Usage

data(electro_explosive)

Format

A data frame with 9 rows and 4 variables:

temp

temperature stress setting in degrees Celsius (35, 45, 55)

it

inspection time in hours (10, 20, 30)

n

number of devices tested at the specified inspection time (10 per row, total 90)

r

number of observed device failures at inspection

Source

Fan, T., Balakrishnan, N., & Chang, C. (2009). The Bayesian approach for highly reliable electro-explosive devices using one-shot device testing. Journal of Statistical Computation and Simulation, 79(9), 1143–1154. doi:10.1080/00949650802142592

References

Zhu, X., Li, Y., Li, T., & Balakrishnan, N. (2026). A simple and efficient eM-algorithm for one-shot device data analysis. Communications in Statistics - Simulation and Computation, 55(3), 959–970. doi:10.1080/03610918.2025.2515193

Examples

data(electro_explosive)
head(electro_explosive)

Control Parameters for EM Algorithms

Description

Sets control parameters for the Expectation-Maximization algorithms.

Usage

em_control(
  maxit = 1000,
  tol = 1e-06,
  trace = FALSE,
  optimizer = c("nlm", "optim"),
  max_m_iter = 100,
  reltol = 1e-10
)

Arguments

Value

A list containing the control options.

Examples

ctrl <- em_control(maxit = 500, tol = 1e-5, trace = FALSE)

Extract Log-Likelihood

Description

Extract Log-Likelihood

Usage

## S3 method for class 'oneshot_em'
logLik(object, ...)

Arguments

Value

An object of class "logLik" representing the maximized log-likelihood value, with additional attributes "df" (degrees of freedom / number of estimated parameters) and "nobs" (total number of tested devices).

Maximum Likelihood Estimation via New eM-Algorithm for One-Shot Device Data

Description

Fits lifetime models for one-shot device data using the new expectation-maximization (eM) algorithm proposed by Zhu et al. (2026).

Usage

oneshot_em(
  formula,
  data,
  it = "it",
  dist = c("exponential", "weibull", "lognormal", "gamma", "custom"),
  sf_fun = NULL,
  theta0 = NULL,
  control = em_control()
)

Arguments

Value

An object of class "oneshot_em" containing:

References

Zhu, X., Li, Y., Li, T., & Balakrishnan, N. (2026). A simple and efficient eM-algorithm for one-shot device data analysis. Communications in Statistics - Simulation and Computation, 55(3), 959–970. doi:10.1080/03610918.2025.2515193

Examples

data(electro_explosive)
fit_exp <- oneshot_em(cbind(r, n) ~ temp, data = electro_explosive,
                      it = "it", dist = "exponential")
summary(fit_exp)

Unified Wrapper for Fitting One-Shot Device ALT Models

Description

Unified high-level fitting interface for one-shot device accelerated life test data.

Usage

oneshot_fit(
  formula,
  data,
  it = "it",
  dist = c("exponential", "weibull", "lognormal", "gamma", "custom"),
  method = c("npm", "tm"),
  sf_fun = NULL,
  theta0 = NULL,
  control = em_control()
)

Arguments

Value

An object of class "oneshot_em" containing model fitting results (see oneshot_em for full list of output elements).

Examples

data(electro_explosive)
fit <- oneshot_fit(cbind(r, n) ~ temp, data = electro_explosive,
                   it = "it", dist = "weibull", method = "npm")
print(fit)

Standardized and Generalized Residuals for One-Shot Device EM Fit

Description

Computes standardized Pearson residuals and randomized quantile residuals for one-shot device data models.

Usage

oneshot_residuals(fit)

Arguments

Value

A data frame containing:

Examples

data(electro_explosive)
fit <- oneshot_em(cbind(r, n) ~ temp, data = electro_explosive, it = "it")
res <- oneshot_residuals(fit)
head(res)

Maximum Likelihood Estimation via Traditional EM-Algorithm for One-Shot Device Data

Description

Fits lifetime models for one-shot device data using the traditional failure-time expectation-maximization (EM) algorithm for comparison with the new eM-algorithm (oneshot_em).

Usage

oneshot_tm(
  formula,
  data,
  it = "it",
  dist = c("exponential", "weibull", "lognormal", "gamma", "custom"),
  sf_fun = NULL,
  theta0 = NULL,
  control = em_control()
)

Arguments

Value

An object of class "oneshot_em" fitted via the traditional EM approach, containing:

References

Balakrishnan, N., & Ling, M. (2012). EM algorithm for one-shot device testing under the exponential distribution. Computational Statistics & Data Analysis, 56(3), 502–509.

Examples

data(electro_explosive)
fit_tm <- oneshot_tm(cbind(r, n) ~ temp, data = electro_explosive,
                     it = "it", dist = "exponential")
summary(fit_tm)

Diagnostic Plots for One-Shot Device EM Fit

Description

Generates fitted survival curves, empirical vs fitted failure proportions, or residual plots.

Usage

## S3 method for class 'oneshot_em'
plot(x, type = c("fitted", "survival", "residuals"), ...)

Arguments

Value

Invisibly returns the original input object x of class "oneshot_em". This function is called for its side effect of producing graphical diagnostic plots (fitted vs observed failure rates, survival curves, or standardized residuals).

Predict Survival Probability or Mean Time To Failure

Description

Predicts survival probability S(t) or Mean Time To Failure (MTTF) under specified stress levels.

Usage

## S3 method for class 'oneshot_em'
predict(object, newdata = NULL, time = NULL, type = c("survival", "mttf"), ...)

Arguments

Value

A numeric vector or matrix of class "numeric" containing predicted survival probabilities (when type = "survival") or Mean Time To Failure (MTTF) values (when type = "mttf") calculated under the specified stress levels and times.

Print Method for One-Shot Device EM Fit

Description

Print Method for One-Shot Device EM Fit

Usage

## S3 method for class 'oneshot_em'
print(x, digits = max(3L, getOption("digits") - 3L), ...)

Arguments

Value

Invisibly returns the original input object x of class "oneshot_em". This function is called primarily for its side effect of printing model fit summary information (distribution, convergence status, estimated coefficients, log-likelihood) to the console.

Print Method for Summary of One-Shot Device EM Fit

Description

Print Method for Summary of One-Shot Device EM Fit

Usage

## S3 method for class 'summary.oneshot_em'
print(x, digits = max(3L, getOption("digits") - 3L), ...)

Arguments

Value

Invisibly returns the summary object x of class "summary.oneshot_em". This function is called primarily for its side effect of formatting and printing model summary results, coefficient matrices, and information criteria to the console.

Summary Method for One-Shot Device EM Fit

Description

Computes summary statistics including standard errors, z-values, p-values, 95% Wald confidence intervals, and model selection criteria.

Usage

## S3 method for class 'oneshot_em'
summary(object, level = 0.95, ...)

Arguments

Value

An object of class "summary.oneshot_em", which is a list containing the following components:

Extract Variance-Covariance Matrix

Description

Extract Variance-Covariance Matrix

Usage

## S3 method for class 'oneshot_em'
vcov(object, ...)

Arguments

Value

A symmetric numeric matrix of class "matrix" representing the estimated variance-covariance matrix of the parameter estimates.