Linear Model with Feature Engineering: Silver Prices Surge

Silver prices have reached a 14-year high amid growing expectations that the US Federal Reserve (FED) will cut interest rates this month.

According to the machine learning model, the bands are down, and the price is above the upper band, indicating anomalous price levels.

Source code:

library(tidyverse)
library(tidymodels)
library(tidyquant)
library(timetk)
library(modeltime)

#Silver Futures
df_silver <- 
  tq_get("SI=F") %>% 
  select(date, close) %>% 
  filter(date >= last(date) - months(36)) %>% 
  drop_na()


#Splitting the data
df_split <- 
  df_silver %>% 
  time_series_split(assess = "30 days", 
                    cumulative = TRUE)

df_train <- 
  training(df_split)

df_test <- 
  testing(df_split)

# Turn the normal mean function into a rolling mean with a 5 row .period
mean_roll_5 <- slidify(mean, .period = 5, .align = "right")

#Preprocessing
rec_spec <- 
  recipe(close ~ ., data = df_train) %>% 
  step_timeseries_signature(date) %>% 
  step_mutate(slid_close = mean_roll_5(close)) %>% 
  step_impute_bag(slid_close) %>% 
  step_fourier(date, period = 365, K = 5) %>%
  step_rm(date) %>%
  step_dummy(all_nominal_predictors(), one_hot = TRUE) %>% 
  step_zv(all_predictors()) %>% 
  step_normalize(all_numeric_predictors())


#Model Specification
mod_spec <- 
  linear_reg() %>% 
  set_engine("lm")


#Training
wflow_fit <- 
  workflow() %>% 
  add_recipe(rec_spec) %>% 
  add_model(mod_spec) %>% 
  fit(df_train)

#Modeltime
df_modeltime <- 
  modeltime_table(wflow_fit)

#Calibrate the model to the testing set
calibration_tbl <- 
  df_modeltime %>%
  modeltime_calibrate(new_data = df_test)


#Accuracy of the finalized model
calibration_tbl %>%
  modeltime_accuracy(metric_set = metric_set(rmse, rsq, mape))

  
#Prediction Intervals
calibration_tbl %>%
  modeltime_forecast(
    new_data    = df_test,
    actual_data = df_test
  ) %>% 
  plot_modeltime_forecast(
    .interactive = FALSE,
    .line_size = 1.5
  )  +
  labs(title = "Silver Futures", 
       subtitle = "<span style = 'color:dimgrey;'>Predictive Intervals</span> of <span style = 'color:red;'>ML Model</span> Model", 
       y = "", x = "") + 
  scale_y_continuous(labels = scales::label_currency()) +
  scale_x_date(labels = scales::label_date("%b %d"),
               date_breaks = "4 days") +
  theme_minimal(base_family = "Roboto Slab", base_size = 16) +
  theme(plot.subtitle = ggtext::element_markdown(face = "bold"),
        plot.title = element_text(face = "bold"),
        plot.background = element_rect(fill = "azure", color = "azure"),
        panel.background = element_rect(fill = "snow", color = "snow"),
        axis.text = element_text(face = "bold"),
        axis.text.x = element_text(angle = 45, 
                                   hjust = 1, 
                                   vjust = 1),
        legend.position = "none")

Related