State any three problems that may arise if data is not prepared properly before training a Machine Learning model.

Three problems that may arise are:

1. If the data has noise or wrong information, the predictions generated by the model will not be correct, and the accuracy of the model will be low.
2. If the collected data is not in a form compatible with the algorithm, the algorithm will not be able to learn the important features from it during training.
3. If missing, duplicate or inconsistent data is not handled during preparation, the model may learn incorrect patterns and produce unreliable predictions.