forecasting inflation via lasso

[richard1999]

I have problems with evaluation lasso model. I am trying to predict inflation and there are still error.

# packages
library(readxl)
library(tidyverse)
library(ggplot2)
library(dplyr)
library(forecast)
library(zoo)
library(glmnet)
library(Metrics)
Canada <- load("Canada_new.RData")

n <- nrow(data)
train_size <- floor(0.7 * n)  # You can adjust the training size as needed

# Create training and test sets
train_data <- head(data, train_size)
test_data <- tail(data, n - train_size)

train_data <- na.omit(train_data)
test_data <- na.omit(test_data)

lasso_model_and_evaluate <- function(data, window_size, horizon) {
    n <- nrow(data)
    mse_values <- c()
    mae_values <- c()

    for (i in 1:(n - window_size - horizon + 1)) {
        train_index <- i:(i + window_size - 1)
        test_index <- (i + window_size):(i + window_size + horizon - 1)

        train_data <- data[train_index, ]
        test_data <- data[test_index, ]

        x_train <- as.matrix(train_data[, -which(names(train_data) == "Inflation")])
        y_train <- train_data$Inflation

        x_test <- as.matrix(test_data[, -which(names(test_data) == "Inflation")])
        y_test <- test_data$Inflation

        lasso_fit <- glmnet(x_train, y_train, alpha = 1)
        predicted_values <- predict(lasso_fit, s = 0.01, newx = x_test)

        mse <- mse(y_test, predicted_values)
        mae <- mae(y_test, predicted_values)

        mse_values <- c(mse_values, mse)
        mae_values <- c(mae_values, mae)
    }

    return(list(mse_values = mse_values, mae_values = mae_values))
}

# Preddefinovanie veľkosti rolujúceho okna
window_size <- 60

# Vykonanie a vyhodnotenie LASSO modelu pre každý horizont
horizons <- c(1, 2, 3, 4)
results <- list()

for (horizon in horizons) {
    result <- lasso_model_and_evaluate(data, window_size, horizon)
    results[[paste("Horizon", horizon)]] <- result
}

# Print out the contents of the `results` list
print(results)

# Check the structure and contents of the results list
str(results)

# Accessing elements of the list
print(results[["Horizon1"]])
print(results[["Horizon2"]])
print(results[["Horizon3"]])
print(results[["Horizon4"]])

Error: $ operator is invalid for atomic vectors

Thank you in advance for any reply.

[nirgrahamuk]

If this accurately reflects your script , the first problem must be that you load data into an object called Canada, and the name is not referred to again. You use an undefined data name.