By Using Statistical Model ( ARIMA ) to Predict the Quantity Demand in future
Raw Data
Checking Trend-Seasonality of Data
- If there is behaviour of increasing or decreasing trend , we have to remove it before modeling
Checking ACF and PACF
- To get possible parameters p and q for ARIMA model
Results of Modeling
Predictions for next three months
- Predicting for next three months and calculate the confidence intervals of predictions
- Confidence Intervels are based on 95% confidence levels and can be considerd as the possible upper and lower bounds.
MAPE between Predctions and Actual Values (3 months)
MAPE(Mean Absolute Percentage Error) : 13.32% / 25.31% /19.55%