ARIMA Modelling and Forecasting Using R - Assignment -17
ARIMA Modelling and Forecasting Using R
Problem Title
Forecasting Monthly Electricity Consumption Using an ARIMA Model
Problem Statement
An electricity distribution company has recorded the following monthly electricity consumption (in thousand units) over three years.
The company wants to analyze the historical consumption pattern and forecast electricity consumption for the next six months using an ARIMA model.
Monthly Electricity Consumption Data
125, 132, 128, 140, 145, 150, 148, 155, 160, 158, 165, 170, 172, 178, 175, 185, 190, 195, 192, 200, 205, 202, 210, 215, 218, 225, 222, 235, 240, 248, 245, 255, 260, 258, 268, 275
Tasks to be Performed by Students
Using R, perform the following:
1. Create a Time Series
Create a monthly time series starting from January 2023.
2. Visualize the Original Time Series
Plot the electricity consumption data and identify whether the data show:
- A trend
- Seasonal variation
- Random fluctuations
3. Check the Need for Differencing
Perform first-order differencing and plot the differenced series.
Compare the original and differenced series.
4. Generate ACF and PACF Plots
Create:
- An ACF plot
- A PACF plot
Use these plots to identify possible values for:
- — AutoRegressive order
- — Moving Average order
Since first-order differencing is performed:
5. Build Candidate ARIMA Models
Based on the ACF and PACF plots, students should try at least three possible models, for example:
6. Compare the Models
Compare the candidate models using:
- AIC
- BIC
Identify the model that provides the best fit.
Generally, a model with lower AIC and BIC values is preferred.
7. Forecast Future Values
Using the selected ARIMA model, forecast the electricity consumption for the next six months.
Display:
- Point forecasts
- 80% prediction interval
- 95% prediction interval
8. Plot the Forecast
Create a graph showing:
- Historical electricity consumption
- Forecasted values
- Prediction intervals
Comments
Post a Comment