| library(shiny) |
| library(DT) |
|
|
| |
| model_data <- data.frame( |
| Methodology = c( |
| "Linear Regression", |
| "Logistic Regression", |
| "Random Forest", |
| "Gradient Boosting (XGBoost)", |
| "Support Vector Machine", |
| "Neural Network (MLP)", |
| "K-Nearest Neighbors", |
| "LASSO Regression", |
| "Decision Tree", |
| "Naive Bayes" |
| ), |
| Category = c( |
| "Regression", "Classification", "Ensemble", "Ensemble", |
| "Kernel-based", "Deep Learning", "Instance-based", |
| "Regression", "Tree-based", "Probabilistic" |
| ), |
| Interpretability = c( |
| "High", "High", "Medium", "Low", |
| "Low", "Low", "Medium", |
| "High", "High", "High" |
| ), |
| Scalability = c( |
| "High", "High", "Medium", "High", |
| "Low", "High", "Low", |
| "High", "Medium", "High" |
| ), |
| Handles_Nonlinearity = c( |
| "No", "No", "Yes", "Yes", |
| "Yes", "Yes", "Yes", |
| "No", "Yes", "No" |
| ), |
| Typical_Use_Case = c( |
| "Continuous outcome prediction", |
| "Binary/multi-class classification", |
| "Tabular data, feature importance", |
| "Competitions, structured data", |
| "High-dimensional classification", |
| "Complex pattern recognition", |
| "Recommendation, anomaly detection", |
| "Feature selection, sparse models", |
| "Simple rule-based decisions", |
| "Text classification, spam filtering" |
| ), |
| Requires_Feature_Scaling = c( |
| "No", "No", "No", "No", |
| "Yes", "Yes", "Yes", |
| "No", "No", "No" |
| ), |
| stringsAsFactors = FALSE |
| ) |
|
|
| |
| ui <- fluidPage( |
| titlePanel("Model / Methodology Comparison"), |
| |
| sidebarLayout( |
| sidebarPanel( |
| width = 3, |
| h4("Filters"), |
| selectInput( |
| "category_filter", "Category:", |
| choices = c("All", sort(unique(model_data$Category))), |
| selected = "All" |
| ), |
| selectInput( |
| "interp_filter", "Interpretability:", |
| choices = c("All", "High", "Medium", "Low"), |
| selected = "All" |
| ), |
| checkboxInput("nonlinear_only", "Only models handling nonlinearity", FALSE), |
| hr(), |
| p("This table summarizes key properties of common modeling methodologies.", |
| style = "color: grey; font-size: 0.9em;") |
| ), |
| |
| mainPanel( |
| width = 9, |
| DTOutput("model_table") |
| ) |
| ) |
| ) |
|
|
| |
| server <- function(input, output, session) { |
| |
| filtered_data <- reactive({ |
| df <- model_data |
| |
| if (input$category_filter != "All") { |
| df <- df[df$Category == input$category_filter, ] |
| } |
| if (input$interp_filter != "All") { |
| df <- df[df$Interpretability == input$interp_filter, ] |
| } |
| if (input$nonlinear_only) { |
| df <- df[df$Handles_Nonlinearity == "Yes", ] |
| } |
| df |
| }) |
| |
| output$model_table <- renderDT({ |
| datatable( |
| filtered_data(), |
| options = list( |
| pageLength = 10, |
| autoWidth = TRUE, |
| columnDefs = list( |
| list(width = "180px", targets = 0), |
| list(width = "200px", targets = 5) |
| ) |
| ), |
| rownames = FALSE, |
| colnames = c( |
| "Methodology", "Category", "Interpretability", "Scalability", |
| "Handles Nonlinearity", "Typical Use Case", "Requires Feature Scaling" |
| ) |
| ) |
| }) |
| } |
|
|
| shinyApp(ui, server) |
