Thanks Jingles.

The following three families of models can serve as good alternatives to the humble yet power Classical Linear Model:

• The GLM family
• The exponential smoothing family
• ARIMA class models

Be sure to not to choose a model based on how well it fits your test data.

Instead, choose a model that is the most appropriate one for the characteristics exhibited by your training data set.

If at all possible, generate a different testing dataset while evaluating each model.

Good luck!