|
| 1 | +##============================================================================= |
| 2 | +#' Variable dependency graph from a uvarpro model |
| 3 | +#' |
| 4 | +#' Extracts cross-variable dependency scores from a fitted \code{uvarpro} |
| 5 | +#' object using \code{\link[varPro]{get.beta.entropy}} and |
| 6 | +#' \code{\link[varPro]{sdependent}}, and returns a tidy list suitable for |
| 7 | +#' \code{plot.gg_udependent}. |
| 8 | +#' |
| 9 | +#' @param object A fitted \code{uvarpro} object (required). |
| 10 | +#' @param threshold Numeric; positive dependency threshold passed to |
| 11 | +#' \code{sdependent()}. An edge \eqn{i \to j} is drawn when |
| 12 | +#' \code{I[i, j] >= threshold}. Default \code{0.25}. |
| 13 | +#' @param q.signal Quantile threshold (0--1) for signal variable selection; |
| 14 | +#' passed to \code{sdependent()}. Default \code{0.75}. |
| 15 | +#' @param directed Logical; \code{TRUE} (default) builds a directed igraph. |
| 16 | +#' @param min.degree Integer or \code{NULL}. When non-\code{NULL}, only nodes |
| 17 | +#' with degree \eqn{\ge} \code{min.degree} are retained in \code{$nodes}, |
| 18 | +#' \code{$edges}, and \code{$graph}. |
| 19 | +#' @param ... Additional arguments forwarded to \code{varPro::sdependent()}. |
| 20 | +#' |
| 21 | +#' @return A named list of class \code{"gg_udependent"} with elements: |
| 22 | +#' \describe{ |
| 23 | +#' \item{\code{$edges}}{Data frame: \code{variable_from}, \code{variable_to}, |
| 24 | +#' \code{weight} (raw cross-importance value).} |
| 25 | +#' \item{\code{$nodes}}{Data frame: \code{variable} (factor, levels by |
| 26 | +#' descending degree), \code{degree} (integer; out-degree when |
| 27 | +#' \code{directed = TRUE}, total degree when \code{directed = FALSE}), |
| 28 | +#' \code{selected} (logical, \code{TRUE} if in \code{sdependent}'s |
| 29 | +#' signal set).} |
| 30 | +#' \item{\code{$graph}}{igraph object. \code{NULL} if no dependencies |
| 31 | +#' detected.} |
| 32 | +#' } |
| 33 | +#' A \code{"provenance"} attribute carries \code{threshold}, \code{q.signal}, |
| 34 | +#' \code{directed}, \code{min.degree}, \code{xvar.names}, and \code{n}. |
| 35 | +#' |
| 36 | +#' @seealso \code{\link{plot.gg_udependent}} |
| 37 | +#' |
| 38 | +#' @examples |
| 39 | +#' \donttest{ |
| 40 | +#' set.seed(42) |
| 41 | +#' uv <- varPro::uvarpro(iris[, -5], ntree = 50) |
| 42 | +#' gg <- gg_udependent(uv) |
| 43 | +#' print(gg) |
| 44 | +#' } |
| 45 | +#' |
| 46 | +#' @importFrom varPro get.beta.entropy sdependent |
| 47 | +#' @importFrom igraph graph_from_adjacency_matrix degree delete_vertices as_data_frame V |
| 48 | +#' @export |
| 49 | +gg_udependent <- function(object, |
| 50 | + threshold = 0.25, |
| 51 | + q.signal = 0.75, |
| 52 | + directed = TRUE, |
| 53 | + min.degree = NULL, |
| 54 | + ...) { |
| 55 | + .validate_udep_inputs(object, threshold, directed) |
| 56 | + |
| 57 | + ## ---- Compute cross-variable dependency matrix ---------------------------- |
| 58 | + imp_mat <- varPro::get.beta.entropy(object) |
| 59 | + |
| 60 | + ## ---- Helper: build and return an empty gg_udependent result --------------- |
| 61 | + .empty_result <- function(msg) { |
| 62 | + warning("gg_udependent: ", msg, |
| 63 | + "\nReturning empty structure. Consider lowering threshold.", |
| 64 | + call. = FALSE) |
| 65 | + empty_edges <- data.frame(variable_from = character(0), |
| 66 | + variable_to = character(0), |
| 67 | + weight = numeric(0), |
| 68 | + stringsAsFactors = FALSE) |
| 69 | + empty_nodes <- data.frame(variable = factor(character(0)), |
| 70 | + degree = integer(0), |
| 71 | + selected = logical(0), |
| 72 | + stringsAsFactors = FALSE) |
| 73 | + result <- structure( |
| 74 | + list(edges = empty_edges, nodes = empty_nodes, graph = NULL), |
| 75 | + class = c("gg_udependent", "list") |
| 76 | + ) |
| 77 | + attr(result, "provenance") <- .udep_provenance(object, threshold, q.signal, |
| 78 | + directed, min.degree) |
| 79 | + result |
| 80 | + } |
| 81 | + |
| 82 | + ## ---- Build adjacency from threshold; short-circuit if empty -------------- |
| 83 | + adj_mat <- (imp_mat >= threshold) * 1 |
| 84 | + diag(adj_mat) <- 0 |
| 85 | + if (sum(adj_mat) == 0L) { |
| 86 | + return(.empty_result( |
| 87 | + paste0("no edges found at threshold=", threshold) |
| 88 | + )) |
| 89 | + } |
| 90 | + |
| 91 | + ## ---- Call sdependent for signal detection -------------------------------- |
| 92 | + sdep <- varPro::sdependent(imp_mat, threshold = threshold, |
| 93 | + q.signal = q.signal, directed = directed, |
| 94 | + min.degree = min.degree, plot = FALSE, ...) |
| 95 | + |
| 96 | + ## ---- Handle empty graph (sdependent may also return character) ----------- |
| 97 | + if (is.character(sdep)) { |
| 98 | + return(.empty_result(sdep)) |
| 99 | + } |
| 100 | + |
| 101 | + ## ---- Build igraph from adjacency ----------------------------------------- |
| 102 | + ## For undirected, symmetrise first so edge existence = max(I[i,j], I[j,i]) |
| 103 | + ## and mode = "undirected" is valid (igraph >= 1.6.0 requires symmetry). |
| 104 | + if (!isTRUE(directed)) { |
| 105 | + adj_mat <- pmax(adj_mat, t(adj_mat)) |
| 106 | + } |
| 107 | + g <- igraph::graph_from_adjacency_matrix( |
| 108 | + adj_mat, |
| 109 | + mode = if (isTRUE(directed)) "directed" else "undirected", |
| 110 | + diag = FALSE |
| 111 | + ) |
| 112 | + isolated <- igraph::degree(g, mode = "all") == 0 |
| 113 | + g <- igraph::delete_vertices(g, which(isolated)) |
| 114 | + |
| 115 | + ## ---- Build tidy edge data frame with raw weights ------------------------- |
| 116 | + edge_df <- igraph::as_data_frame(g, what = "edges") |
| 117 | + if (nrow(edge_df) > 0L) { |
| 118 | + if (isTRUE(directed)) { |
| 119 | + edge_df$weight <- mapply(function(i, j) imp_mat[i, j], |
| 120 | + edge_df[[1L]], edge_df[[2L]]) |
| 121 | + } else { |
| 122 | + ## Undirected: weight = max of both directions |
| 123 | + edge_df$weight <- mapply( |
| 124 | + function(i, j) max(imp_mat[i, j], imp_mat[j, i]), |
| 125 | + edge_df[[1L]], edge_df[[2L]]) |
| 126 | + } |
| 127 | + } else { |
| 128 | + edge_df$weight <- numeric(0) |
| 129 | + } |
| 130 | + names(edge_df)[1:2] <- c("variable_from", "variable_to") |
| 131 | + |
| 132 | + ## ---- Build tidy node data frame ------------------------------------------ |
| 133 | + vnames <- igraph::V(g)$name |
| 134 | + ## degree: out-degree for directed (matches sdependent's signal.vars logic), |
| 135 | + ## total degree for undirected |
| 136 | + deg_vec <- if (isTRUE(directed)) { |
| 137 | + igraph::degree(g, mode = "out")[vnames] |
| 138 | + } else { |
| 139 | + igraph::degree(g)[vnames] |
| 140 | + } |
| 141 | + |
| 142 | + signal_set <- if (is.null(sdep$signal.vars)) character(0) else sdep$signal.vars |
| 143 | + node_df <- data.frame( |
| 144 | + variable = factor(vnames, levels = vnames[order(-deg_vec)]), |
| 145 | + degree = as.integer(deg_vec), |
| 146 | + selected = vnames %in% signal_set, |
| 147 | + stringsAsFactors = FALSE, |
| 148 | + row.names = NULL |
| 149 | + ) |
| 150 | + |
| 151 | + ## ---- Apply min.degree node filtering (user-requested subsetting) --------- |
| 152 | + if (!is.null(min.degree)) { |
| 153 | + keep <- node_df$degree >= min.degree |
| 154 | + keep_names <- as.character(node_df$variable)[keep] |
| 155 | + drop_names <- as.character(node_df$variable)[!keep] |
| 156 | + g <- igraph::delete_vertices(g, drop_names) |
| 157 | + edge_df <- edge_df[ |
| 158 | + edge_df$variable_from %in% keep_names & |
| 159 | + edge_df$variable_to %in% keep_names, , drop = FALSE] |
| 160 | + node_df <- node_df[keep, , drop = FALSE] |
| 161 | + rownames(edge_df) <- NULL |
| 162 | + rownames(node_df) <- NULL |
| 163 | + } |
| 164 | + |
| 165 | + ## ---- Set igraph node attributes ------------------------------------------ |
| 166 | + if (length(igraph::V(g)) > 0L) { |
| 167 | + igraph::V(g)$degree <- node_df$degree[ |
| 168 | + match(igraph::V(g)$name, as.character(node_df$variable))] |
| 169 | + igraph::V(g)$selected <- node_df$selected[ |
| 170 | + match(igraph::V(g)$name, as.character(node_df$variable))] |
| 171 | + } |
| 172 | + |
| 173 | + ## ---- Set igraph edge weights (order-insensitive for undirected) ----------- |
| 174 | + if (length(igraph::E(g)) > 0L && nrow(edge_df) > 0L) { |
| 175 | + el <- igraph::as_data_frame(g, what = "edges") |
| 176 | + if (isTRUE(directed)) { |
| 177 | + idx <- match(paste(el$from, el$to), |
| 178 | + paste(edge_df$variable_from, edge_df$variable_to)) |
| 179 | + } else { |
| 180 | + key_g <- paste(pmin(el$from, el$to), pmax(el$from, el$to)) |
| 181 | + key_e <- paste(pmin(edge_df$variable_from, edge_df$variable_to), |
| 182 | + pmax(edge_df$variable_from, edge_df$variable_to)) |
| 183 | + idx <- match(key_g, key_e) |
| 184 | + } |
| 185 | + igraph::E(g)$weight <- edge_df$weight[idx] |
| 186 | + } |
| 187 | + |
| 188 | + ## ---- Assemble result ------------------------------------------------------ |
| 189 | + result <- structure( |
| 190 | + list(edges = edge_df, nodes = node_df, graph = g), |
| 191 | + class = c("gg_udependent", "list") |
| 192 | + ) |
| 193 | + attr(result, "provenance") <- .udep_provenance(object, threshold, q.signal, |
| 194 | + directed, min.degree) |
| 195 | + result |
| 196 | +} |
| 197 | + |
| 198 | +## ---- Internal helpers ------------------------------------------------------- |
| 199 | + |
| 200 | +#' @keywords internal |
| 201 | +.validate_udep_inputs <- function(object, threshold, directed) { |
| 202 | + if (missing(object) || is.null(object)) { |
| 203 | + stop("'object' must be a fitted uvarpro object.", call. = FALSE) |
| 204 | + } |
| 205 | + if (!inherits(object, "uvarpro")) { |
| 206 | + stop("'object' must be a uvarpro fit (class \"uvarpro\").", call. = FALSE) |
| 207 | + } |
| 208 | + if (!is.numeric(threshold) || length(threshold) != 1L || threshold <= 0) { |
| 209 | + stop("'threshold' must be a single positive numeric value.", call. = FALSE) |
| 210 | + } |
| 211 | + if (!is.logical(directed) || length(directed) != 1L) { |
| 212 | + stop("'directed' must be a single logical value.", call. = FALSE) |
| 213 | + } |
| 214 | + invisible(NULL) |
| 215 | +} |
| 216 | + |
| 217 | +#' @keywords internal |
| 218 | +.udep_provenance <- function(object, threshold, q.signal, directed, min.degree) { |
| 219 | + list( |
| 220 | + threshold = threshold, |
| 221 | + q.signal = q.signal, |
| 222 | + directed = directed, |
| 223 | + min.degree = min.degree, |
| 224 | + xvar.names = object$xvar.names, |
| 225 | + n = nrow(object$x) |
| 226 | + ) |
| 227 | +} |
0 commit comments