Shiny吗啡预测应用报错：缺少sidebar参数无默认值，寻求修复方案

错误原因 #

dashboardPage是shinydashboard的核心布局函数，必须传入header、sidebar、body三个必填参数。你的原代码仅传入了dashboardHeader，dashboardSidebar和dashboardBody被写成了独立语句，未作为参数传入dashboardPage，导致抛出"缺少sidebar参数"的错误。此外，侧边栏和标签页的嵌套结构也存在错误：sidebarMenu需放在dashboardSidebar内部，tabItem必须嵌套在tabItems中再放入dashboardBody。

修正后的UI代码 #

# 加载依赖包
library(shiny)
library(shinydashboard)
library(ggplot2)
library(dplyr)
library(randomForest)
library(Metrics)

# 构建UI
ui <- dashboardPage(
  # 页面头部
  dashboardHeader(title = '吗啡消耗量预测工具', titleWidth = 290),
  
  # 侧边栏
  dashboardSidebar(
    sidebarMenu(
      id = "menu",
      menuItem("变量可视化", tabName = "plots", icon = icon('poll')),
      menuItem("数据看板", tabName = "dash", icon = icon('tachometer-alt')),
      menuItem("消耗量预测", tabName = "pred", icon = icon('search'))
    )
  ),
  
  # 主内容区
  dashboardBody(
    tags$head(tags$style(HTML('.main-header .logo {font-weight: bold;}'))),
    tabItems(
      # 可视化标签页
      tabItem('plots', 
              box(status = 'primary', title = '直方图筛选变量', 
                  selectInput('num', "数值型变量:", c("Age", "BMI", "IV_Fluids", "Operative_times", "Blood_loss", "Time_to_Aldrete_9", "morphine_consumption_24h1",
                                                     "VAS_basalR", "VAS_basalM", "VAS_2hrR", "VAS_2hrM", "VAS_4hrM", "VAS-4hrR",
                                                     "VAS_8hrR", "VAS_8hrM", "VAS_12hrR", "VAS_12hrM", "VAS_16hrR", "VAS_16hrM", 
                                                     "VAS_24hrR", "VAS_24hrM", "QOR_psychological_support", "QOR_emotional_state",
                                                     "QOR_Physical_comfort", "QOR_physical_independence", "QOR_Pain", "Total")),
                  footer = '数值变量分布直方图'),
              box(status = 'primary', title = '频数图筛选变量',
                  selectInput('cat', '分类型变量:', c("ASA", "Postoperative_vomiting", "Sedation_0to8h", "Sedation_9to16h", "Sedation_17to24h")),
                  footer = '分类变量频数统计图'),
              box(plotOutput('histPlot')),
              box(plotOutput('freqPlot'))
      ),
      
      # 预测标签页
      tabItem('pred',
              box(title = '分类变量输入', 
                  status = 'primary', width = 12, 
                  splitLayout(
                    tags$head(tags$style(HTML(".shiny-split-layout > div {overflow: visible;}"))),
                    cellWidths = c('19%', '4%', '19%', '4%', '19%', '4%', '19%', '4%', '8%'),
                    selectInput( 'p_group', '分组', c("0", "30", "60", "90")),
                    div(),
                    selectInput('p_ASA', 'ASA分级', c('1', '2', '3')),
                    div(),
                    selectInput( 'p_Sedation_17to24h', '17-24h镇静评分', c('1', '2', '3', '4')),
                    div(),
                    radioButtons( 'p_Postoperative_vomiting', '术后呕吐', c('是', '否'))
                  )
              ),
              box(title = '数值变量输入',
                  status = 'primary', width = 12,
                  splitLayout(cellWidths = c('22%', '4%','21%', '4%', '21%', '4%', '21%'),
                              sliderInput( 'p_Age', '年龄(岁)', min = 0, max = 100, value = 50),
                              div(),
                              numericInput( 'p_BMI', 'BMI指数', 22),
                              div(),
                              numericInput( 'p_VAS_24hrM', '24h活动时VAS评分', 3),
                              div(),
                              numericInput( 'p_QOR_psychological_support', 'QOR-心理支持评分', 10),
                              div(),
                              numericInput( 'p_QOR_Pain', 'QOR-疼痛评分', 10),
                              numericInput( 'p_QOR_Physical_comfort', 'QOR-躯体舒适度评分', 10)
                  )
              ),
              box(title = '预测结果',
                  status = 'success', 
                  solidHeader = TRUE, 
                  width = 4, height = 260,
                  div(h5('吗啡24小时消耗量(mg):')),
                  verbatimTextOutput("value", placeholder = TRUE),
                  div(h5('消耗量预测范围:')),
                  verbatimTextOutput("range", placeholder = TRUE),
                  actionButton('cal','计算', icon = icon('calculator'))
              ),
              box(title = '模型说明',
                  status = 'success', 
                  width = 8, height = 260,
                  helpText('本模型基于年龄、BMI、24h活动时VAS评分、术后恢复质量(QOR)等变量，预测患者24小时吗啡消耗量。'),
                  helpText('训练数据集来自UCI机器学习仓库，研究主题为"术前度洛西汀对改良根治术患者的短期疗效"，包含17379条观测记录。'),
                  helpText(sprintf('模型采用随机森林算法，训练集上的平均绝对误差(MAE)为%s mg，均方根误差(RMSE)为%s mg。', round(mae_rf, digits = 0), round(rmse_rf, digits = 0)))
              )
      )
    )
  )
)

修正后的Server代码 #

server <- shinyServer(function(input, output) {
  
  # 变量可视化模块
  output$histPlot <- renderPlot({
    # 补充你的直方图绘制逻辑，示例：
    # data %>% ggplot(aes_string(x = input$num)) + geom_histogram(bins = 15, fill = "#2c3e50") + theme_minimal()
  })
  
  output$freqPlot <- renderPlot({
    # 补充你的频数图绘制逻辑，示例：
    # data %>% ggplot(aes_string(x = input$cat)) + geom_bar(fill = "#3498db") + theme_minimal()
  })
  
  # 数据看板模块（按需补充）
  output$linePlot <- renderPlot({})
  output$barPlot <- renderPlot({})
  
  # 预测模块
  a <- reactiveValues(result = NULL)
  
  observeEvent(input$cal, {
    # 注意：需提前加载训练好的模型model_rf、测试集test_set、误差指标mae_rf
    # 移除原代码中未定义的p_mnth参数
    values = data.frame(
      Group = input$p_group,
      ASA = input$p_ASA,
      Sedation_17to24hr = input$p_Sedation_17to24h,
      PONV = input$p_Postoperative_vomiting,
      Age = input$p_Age,
      BMI = input$p_BMI,
      VAS_24hrM = input$p_VAS_24hrM,
      QOR_psychological_support = input$p_QOR_psychological_support,
      QOR_Pain = input$p_QOR_Pain,
      QOR_Physical_comfort = input$p_QOR_Physical_comfort
    )
    
    # 确保输入变量名与模型训练时的特征名完全一致
    a$result <-  round(predict(model_rf, newdata = values), digits = 0)
  })
  
  output$value <- renderText({
    if(is.null(a$result)) "请点击计算按钮" else paste(a$result)
  })
  
  output$range <- renderText({
    input$cal
    isolate({
      if(is.null(a$result)) "" else sprintf('%s - %s mg', 
                                            round(a$result - mae_rf, digits = 0),
                                            round(a$result + mae_rf, digits = 0))
    })
  })
  
})

shinyApp(ui, server)

其他注意事项 #

1. 前置依赖准备：需提前加载训练好的随机森林模型model_rf、测试集test_set、误差指标mae_rf和rmse_rf，确保这些对象在运行Shiny应用前已定义。
2. 变量名一致性：输入控件的命名需与模型训练时的特征名完全匹配，否则预测会失败。
3. 可视化逻辑补充：需在renderPlot中补充具体的绘图代码，基于你使用的数据集完成变量可视化功能。

内容的提问来源于stack exchange，提问作者Aly