>Since you have a literal dependency between one data point and the next, you can't train your model using randomized data This is why the train/test split was

>Since you have a literal dependency between one data point and the next, you can't train your model using randomized data This is why the train/test split was not randomized, but sequential.