You are working on a classification problem involving time series data. Early experiments have yielded an AUC ROC value of 99% on the training set, even though you have not used any advanced algorithms or hyperparameter tuning. Given these results, what should be your next step to diagnose and address potential issues with your model?