如何获取R中拟合函数的名称

Question

我构建了一个代码，可以将不同的联结类型与数据相匹配，并根据最小的 AIC 值选择最佳的联结类型。该代码运行良好。然而结果如下：

Normal copula, dim. d = 3 
Dimension:  3 
Parameters:
  rho.1   = 0.8094
dispstr:  ex 

然后我可以从结果中获取维度和参数，但可以获取联结的名称。也就是说，我需要提取 copual 的名称。例如，

    selected_copula@parameters
[1] 0.7833736

我也希望能够使用

selected_copula@copula

来获取最适合的联结函数的名称，但它不起作用。

fitCopula

函数允许我们使用

@copula

提取联结的名称，但我无法使用我的代码来提取联结的名称。

完整代码如下：

library(copula)
fit_select_copula <- function(copula_list, data) {
  fit_copula_models <- function(copula_list, data) {
    num_copulas <- length(copula_list)
    copula_results <- vector("list", length = num_copulas)

    for (i in seq_along(copula_list)) {
      copula <- copula_list[[i]]
      fitted_model <- fitCopula(copula,data)

      copula_results[[i]] <- list(
        copula = fitted_model@copula,
        loglik = fitted_model@loglik,
        aic = AIC(fitted_model)
      )
    }

    return(copula_results)
  }

  copula_results <- fit_copula_models(copula_list, data)

  aic_values <- sapply(copula_results, function(x) x$aic)
  lowest_aic_index <- which.min(aic_values)

  selected_copula <- copula_results[[lowest_aic_index]]$copula

  return(selected_copula)
}

# Example usage
data <- rCopula(1000, normalCopula(0.8, dim = 3))  # Replace with your actual data
copula_list <- list(
  frank = frankCopula(6, 3),
  clayton = claytonCopula(2, 3),
  normal = normalCopula(0.7, 3)
)

selected_copula <- fit_select_copula(copula_list, data)

Answer 1

我想如果你想在你的初始名单中使用你自己的名字；然后，您需要添加足够的簿记以在您的运营中携带这些内容。

library(copula)
fit_select_copula <- function(copula_list, data) {
  fit_copula_models <- function(copula_list, data) {
    num_copulas <- length(copula_list)
    copula_names <- names(copula_list) #  <<< ADDED THIS  <<<<
    copula_results <- vector("list", length = num_copulas)
    
    for (i in seq_along(copula_list)) {
      copula <- copula_list[[i]]
      fitted_model <- fitCopula(copula,data)
      
      copula_results[[i]] <- list(
        copula_name = copula_names[[i]], #  <<< ADDED THIS  <<<<
        copula = fitted_model@copula,
        loglik = fitted_model@loglik,
        aic = AIC(fitted_model)
      )
    }
    
    return(copula_results)
  }
  
  copula_results <- fit_copula_models(copula_list, data)
  aic_values <- sapply(copula_results, function(x) x$aic)
  lowest_aic_index <- which.min(aic_values)
  
 ### ALTERED THIS ####
  selected_copula <- list(name=copula_results[[lowest_aic_index]]$copula_name,
                          content = copula_results[[lowest_aic_index]]$copula)
  
  return(selected_copula)
}

# Example usage
data <- rCopula(1000, normalCopula(0.8, dim = 3))  # Replace with your actual data
copula_list <- list(
  frank = frankCopula(6, 3),
  clayton = claytonCopula(2, 3),
  normal = normalCopula(0.7, 3)
)

selected_copula <- fit_select_copula(copula_list, data)