R/print.R

In densityratio: Distribution Comparison Through Density Ratio Estimation

#' Print a \code{ulsif} object
#'
#' @rdname print.ulsif
#' @method print ulsif
#' @param x Object of class \code{ulsif}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{ulsif} object.
#' @importFrom utils str
#' @export
#' @seealso \code{\link{print}}, \code{\link{ulsif}}
#' @example inst/examples/ulsif-example.R



print.ulsif <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )
  cat("  sigma:")
  cat(str(unname(x$sigma)))
  cat("\nRegularization parameter (lambda):")
  cat(str(unname(x$lambda)))
  cat("\nOptimal sigma (loocv): ", paste(format(x$sigma_opt, digits, ...)), "\n",
    "Optimal lambda (loocv): ", paste(format(x$lambda_opt, digits, ...)), "\n",
    sep = ""
  )
  cat("Optimal kernel weights (loocv):")
  cat(str(unname(x$alpha_opt)), "\n")
  invisible(x)
}

#' Print a \code{summary.ulsif} object
#'
#' @rdname print.summary.ulsif
#' @method print summary.ulsif
#' @param x Object of class \code{summary.ulsif}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{summary.ulsif} object.
#' @importFrom utils str
#' @seealso \code{\link{print}}, \code{\link{summary.ulsif}}, \code{\link{ulsif}}
#'
#' @export
#' @example inst/examples/ulsif-example.R


print.summary.ulsif <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )

  cat("\nOptimal sigma: ", paste(format(x$sigma_opt, digits, ...)), "\n",
    "Optimal lambda: ", paste(format(x$lambda_opt, digits, ...)), "\n",
    sep = ""
  )
  cat("Optimal kernel weights:")
  cat(str(unname(x$alpha_opt)), "\n")
  cat("Pearson divergence between P(nu) and P(de): ", paste(format(x$PE, digits = digits, ...)), "\n", sep = "")
  if (!is.null(x$p_value)) {
    cat("Pr(P(nu)=P(de))",
      ifelse(x$p_value < 0.001,
        paste(" < .001"),
        paste(" = ", format(x$p_value, digits = 3, ...))
      ),
      "\nBonferroni-corrected for testing with r(x) = P(nu)/P(de) AND r*(x) = P(de)/P(nu).",
      "\n\n",
      sep = ""
    )
  } else {
    cat("For a two-sample homogeneity test, use 'summary(x, test = TRUE)'.\n\n")
  }
  invisible(x)
}

#' Print a \code{kliep} object
#'
#' @rdname print.kliep
#' @method print kliep
#' @param x Object of class \code{kliep}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{kliep} object.
#' @importFrom utils str
#' @export
#' @seealso \code{\link{print}}, \code{\link{kliep}}
#' @example inst/examples/kliep-example.R


print.kliep <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )
  cat("  sigma:")
  cat(str(unname(x$sigma)))
  if (!is.null(x$cv_score)) {
    cat("\nOptimal sigma (", paste(x$nfold), "-fold cv): ", paste(format(x$sigma_opt, digits = digits, ...), collapse = "  "), "\n", sep = "")
    cat("Optimal kernel weights (", paste(x$nfold), "-fold cv): ", sep = "")
    cat(str(unname(x$alpha_opt)))
  } else {
    cat("\nOptimal sigma: NULL (no cross-validation)\n", sep = "")
    cat("Optimal kernel weights: NULL (no cross-validation)\n", sep = "")
  }
  cat("\nOptimization parameters:\n", sep = "")
  cat("  Learning rate (epsilon): ", paste(format(x$epsilon, digits = digits, ...), collapse = "  "), "\n", sep = "")
  cat("  Maximum number of iterations: ", paste(x$maxit, collapse = "  "))
  cat("\n")
  invisible(x)
}

#' Print a \code{summary.kliep} object
#'
#' @rdname print.summary.kliep
#' @method print summary.kliep
#' @param x Object of class \code{summary.kliep}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{summary.kliep} object.
#' @importFrom utils str
#' @seealso \code{\link{print}}, \code{\link{summary.kliep}}, \code{\link{kliep}}
#'
#' @export
#' @example inst/examples/kliep-example.R


print.summary.kliep <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )

  cat("Optimal sigma: ", paste(format(x$sigma_opt, digits, ...)), "\n", sep = "")
  cat("Optimal kernel weights:")
  cat(str(unname(x$alpha_opt)), "\n")
  cat("Kullback-Leibler divergence between P(nu) and P(de): ", paste(format(x$UKL, digits = digits, ...)), "\n", sep = "")
  if (!is.null(x$p_value)) {
    cat("Pr(P(nu)=P(de))",
      ifelse(x$p_value < 0.001,
        paste(" < .001"),
        paste(" = ", format(x$p_value, digits = 3, ...))
      ),
      "\nBonferroni-corrected for testing with r(x) = P(nu)/P(de) AND r*(x) = P(de)/P(nu).",
      "\n\n",
      sep = ""
    )
  } else {
    cat("For a two-sample homogeneity test, use 'summary(x, test = TRUE)'.\n\n")
  }
  invisible(x)
}

#' Print a \code{kmm} object
#'
#' @rdname print.kmm
#' @method print kmm
#' @param x Object of class \code{kmm}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{kmm} object.
#' @importFrom utils str
#' @export
#' @seealso \code{\link{print}}, \code{\link{kmm}}
#' @example inst/examples/kmm-example.R


print.kmm <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )
  cat("  sigma:")
  cat(str(unname(x$sigma)))
  if (!is.null(x$cv_score)) {
    cat("\nOptimal sigma (", paste(x$nfold), "-fold cv): ", paste(format(x$sigma_opt, digits = digits, ...), collapse = "  "), "\n", sep = "")
    cat("Optimal kernel weights (", paste(x$nfold), "-fold cv): ", sep = "")
    cat(str(unname(x$alpha_opt)))
  } else {
    cat("\nOptimal sigma: NULL (no cross-validation)\n", sep = "")
    cat("Optimal kernel weights: NULL (no cross-validation)\n", sep = "")
  }
  cat("\nOptimization parameters:\n", sep = "")
  cat("  Optimization method: ", ifelse(x$constrained, "Constrained", "Unconstrained"), "\n")
  cat("\n")
  invisible(x)
}

#' Print a \code{summary.kmm} object
#'
#' @rdname print.summary.kmm
#' @method print summary.kmm
#' @param x Object of class \code{summary.kmm}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{summary.kmm} object.
#' @importFrom utils str
#' @seealso \code{\link{print}}, \code{\link{summary.kmm}}, \code{\link{kmm}}
#'
#' @export
#' @example inst/examples/kmm-example.R


print.summary.kmm <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )

  cat("Optimal sigma: ", paste(format(x$sigma_opt, digits, ...)), "\n", sep = "")
  cat("Optimal kernel weights:")
  cat(str(unname(x$alpha_opt)), "\n")
  cat("Pearson divergence between P(nu) and P(de): ", paste(format(x$PE, digits = digits, ...)), "\n", sep = "")
  if (!is.null(x$p_value)) {
    cat("Pr(P(nu)=P(de))",
      ifelse(x$p_value < 0.001,
        paste(" < .001"),
        paste(" = ", format(x$p_value, digits = 3, ...))
      ),
      "\nBonferroni-corrected for testing with r(x) = P(nu)/P(de) AND r*(x) = P(de)/P(nu).",
      "\n\n",
      sep = ""
    )
  } else {
    cat("For a two-sample homogeneity test, use 'summary(x, test = TRUE)'.\n\n")
  }
  invisible(x)
}

#' Print a \code{lhss} object
#'
#' @rdname print.lhss
#' @method print lhss
#' @param x Object of class \code{lhss}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{lhss} object.
#' @importFrom utils str
#' @export
#' @seealso \code{\link{print}}, \code{\link{lhss}}
#' @example inst/examples/lhss-example.R

print.lhss <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )
  cat("  sigma:")
  cat(str(unname(x$sigma)))
  cat("\nRegularization parameter (lambda):")
  cat(str(unname(x$lambda)))
  cat("\nSubspace dimension (m): ", x$m, "\n", sep = "")
  cat("Optimal sigma: ", paste(format(x$sigma_opt, digits, ...)), "\n",
    "Optimal lambda: ", paste(format(x$lambda_opt, digits, ...)), "\n",
    sep = ""
  )
  cat("Optimal kernel weights (loocv):")
  cat(str(unname(x$alpha_opt)), "\n")
  invisible(x)
}

#' Print a \code{summary.lhss} object
#'
#' @rdname print.summary.lhss
#' @method print summary.lhss
#' @param x Object of class \code{summary.lhss}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{summary.lhss} object.
#' @importFrom utils str
#' @seealso \code{\link{print}}, \code{\link{summary.lhss}}, \code{\link{lhss}}
#'
#' @export
#' @example inst/examples/lhss-example.R

print.summary.lhss <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )
  cat("\nSubspace dimension (m): ", x$m, "\n", sep = "")
  cat("Optimal sigma: ", paste(format(x$sigma_opt, digits, ...)), "\n",
    "Optimal lambda: ", paste(format(x$lambda_opt, digits, ...)), "\n",
    sep = ""
  )
  cat("Optimal kernel weights (loocv):")
  cat(str(unname(x$alpha_opt)), "\n")
  cat("Pearson divergence between P(nu) and P(de): ", paste(format(x$PE, digits = digits, ...)), "\n", sep = "")
  if (!is.null(x$p_value)) {
    cat("Pr(P(nu)=P(de))",
      ifelse(x$p_value < 0.001,
        paste(" < .001"),
        paste(" = ", format(x$p_value, digits = 3, ...))
      ),
      "\nBonferroni-corrected for testing with r(x) = P(nu)/P(de) AND r*(x) = P(de)/P(nu).",
      "\n\n",
      sep = ""
    )
  } else {
    cat("For a two-sample homogeneity test, use 'summary(x, test = TRUE)'.\n\n")
  }
  invisible(x)
}

#' Print a \code{spectral} object
#'
#' @rdname print.spectral
#' @method print spectral
#' @param x Object of class \code{spectral}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{spectral} object.
#' @importFrom utils str
#' @export
#' @seealso \code{\link{print}}, \code{\link{spectral}}
#' @example inst/examples/spectral-example.R



print.spectral <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )
  cat("  sigma:")
  cat(str(unname(x$sigma)))
  cat("\n")
  cat("\nSubspace dimension (J):")
  cat(str(unname(x$m)))
  cat("\nOptimal sigma: ", paste(format(x$sigma_opt, digits, ...)), "\n",
    "Optimal subspace: ", paste(format(x$m_opt, digits, ...)), "\n",
    sep = ""
  )
  cat("Optimal kernel weights (cv):")
  cat(str(unname(x$alpha_opt)), "\n")
  invisible(x)
}

#' Print a \code{summary.spectral} object
#'
#' @rdname print.summary.spectral
#' @method print summary.spectral
#' @param x Object of class \code{summary.spectral}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{summary.spectral} object.
#' @importFrom utils str
#' @seealso \code{\link{print}}, \code{\link{summary.spectral}}, \code{\link{spectral}}
#'
#' @export
#' @example inst/examples/spectral-example.R


print.summary.spectral <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Kernel Information:\n",
    "  Kernel type: Gaussian with L2 norm distances\n",
    "  Number of kernels: ", paste0(nrow(x$centers)), "\n",
    sep = ""
  )

  cat("\nOptimal sigma: ", paste(format(x$sigma_opt, digits, ...)), "\n",
    "Optimal subspace: ", paste(format(x$m_opt, digits, ...)), "\n",
    sep = ""
  )
  cat("Optimal kernel weights (cv):")
  cat(str(unname(x$alpha_opt)), "\n")
  cat("Pearson divergence between P(nu) and P(de): ", paste(format(x$PE, digits = digits, ...)), "\n", sep = "")
  if (!is.null(x$p_value)) {
    cat("Pr(P(nu)=P(de))",
      ifelse(x$p_value < 0.001,
        paste(" < .001"),
        paste(" = ", format(x$p_value, digits = 3, ...))
      ),
      "\nBonferroni-corrected for testing with r(x) = P(nu)/P(de) AND r*(x) = P(de)/P(nu).",
      "\n\n",
      sep = ""
    )
  } else {
    cat("For a two-sample homogeneity test, use 'summary(x, test = TRUE)'.\n\n")
  }
  invisible(x)
}

#' Print a \code{naivedensityratio} object
#'
#' @rdname print.naivedensityratio
#' @method print naivedensityratio
#' @param x Object of class \code{naivesubspacedensityratio}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{naivedensityratio} object.
#' @importFrom utils str
#' @export
#' @seealso \code{\link{print}}, \code{\link{naive}}
#' @example inst/examples/naive-example.R


print.naivedensityratio <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Naive density ratio\n",
    "  Number of variables: ", ncol(x$model_matrices$nu), "\n",
    "  Number of numerator samples: ", nrow(x$model_matrices$nu), "\n",
    "  Number of denominator samples: ", nrow(x$model_matrices$de), "\n",
    sep = ""
  )
  cat("  Numerator density:")
  cat(str(unname(stats::predict(x, newdata = x$df_numerator))))
  cat("  Denominator density:")
  cat(str(unname(stats::predict(x, newdata = x$df_denominator))), "\n")

  invisible(x)
}


#' Print a \code{summary.naivedensityratio} object
#'
#' @rdname print.summary.naivedensityratio
#' @method print summary.naivedensityratio
#' @param x Object of class \code{summary.naivedensityratio}.
#' @param digits Number of digits to use when printing the output.
#' @param ... further arguments on how to format the number of digits.
#' @return \code{invisble} The inputted \code{summary.naivedensityratio} object.
#' @importFrom utils str
#' @seealso \code{\link{print}}, \code{\link{summary.naivedensityratio}},
#' \code{\link{naive}}
#'
#' @export
#' @example inst/examples/naive-example.R


print.summary.naivedensityratio <- function(x, digits = max(3L, getOption("digits") - 3L), ...) {
  cat("\nCall:\n", paste0(deparse(x$call)), "\n", sep = "")
  cat("\n")
  cat("Naive density ratio estimate:\n",
    "  Number of variables: ", x$nvars, "\n",
    "  Number of numerator samples: ", x$n[1], "\n",
    "  Number of denominator samples: ", x$n[2], "\n",
    sep = ""
  )
  cat("  Density ratio for numerator samples:")
  cat(str(unname(x$dr$dr[1:x$n[1]])))
  cat("  Density ratio for denominator samples:")
  cat(str(unname(x$dr$dr[(x$n[1] + 1):(x$n[1] + x$n[2])])), "\n\n")

  cat("Squared average log density ratio difference for numerator and denominator samples (SALDRD): ",
    paste(format(x$SALDRD, digits = digits, ...)), "\n",
    sep = ""
  )
  if (!is.null(x$p_value)) {
    cat("Pr(P(nu)=P(de))",
      ifelse(x$p_value < 0.001,
        paste(" < .001"),
        paste(" = ", format(x$p_value, digits = 3, ...))
      ),
      "\n\n",
      sep = ""
    )
  } else {
    cat("For a two-sample homogeneity test, use 'summary(x, test = TRUE)'.\n\n")
  }
  invisible(x)
}