As a data scientists, I want to make sure that all available data is used to train a model and verify our initial results.
The method employed here is rather slow and compute intensive, therefore we trained models on ~10% of the data to quickly iterate on our initial experiments. This small sample size lead to poor performance results. Before exploring additional data management or alternative learning algorithms, we want to rerun the training on the entire data set to confirm our conclusions.
Acceptance Criteria:
[x] Update training_experiments notebook with results from entire dataset.
MichaelClifford
commented
3 years ago
As a data scientists, I want to make sure that all available data is used to train a model and verify our initial results.
The method employed here is rather slow and compute intensive, therefore we trained models on ~10% of the data to quickly iterate on our initial experiments. This small sample size lead to poor performance results. Before exploring additional data management or alternative learning algorithms, we want to rerun the training on the entire data set to confirm our conclusions.
Acceptance Criteria: