For AI project efficiency the data needs to be relevant and authentic.
If it is not relevant and authentic then the testing data may go wrong or produced irrelevant predictions.
For example, while processing data for a cricket match for net run rate prediction, the testing data should be provided for batting, not bowling. If the bowler’s data will be processed as testing data then the net run rate will be predicted in the wrong direction.
Hence, for any AI project to be efficient, the training data should be authentic and relevant to the problem statement scoped.
Study more about Data Acquisition at Data Acquisition Class 10