You built a demand forecasting model for a grocery delivery service. The service experienced a surge in orders during holiday seasons, and the model's predictions are now inaccurate due to these periodic fluctuations. What steps should you take to improve the model's accuracy and avoid similar issues in the future?