datasetsDynamic 0.0.5

A package to load datasets for benchmarking prescriptive analytics approaches dynamically

datasetsDynamic

Install

pip install datasetsDynamic

How to use

For every dataset a load function is implemented which computes training and test data for the corresponding dataset including all preprocessing and basic feature engineering steps. For most datasets the test period can be chosen dynamically using the parameter testDays. While doing so, it is ensured that all features that depend on the train and test structure are computed only based on the training data.

from datasetsDynamic.loadDataYaz import loadDataYaz
data, XTrain, yTrain, XTest, yTest = loadDataYaz(testDays = 28, returnXY = True, daysToCut = 0, disable_progressbar = False)

Rolling: 100%|██████████| 30/30 [00:00<00:00, 36.35it/s]
Feature Extraction: 100%|██████████| 30/30 [00:02<00:00, 13.59it/s]
Rolling: 100%|██████████| 30/30 [00:00<00:00, 35.29it/s]
Feature Extraction: 100%|██████████| 30/30 [00:02<00:00, 12.19it/s]
Rolling: 100%|██████████| 30/30 [00:00<00:00, 37.20it/s]
Feature Extraction: 100%|██████████| 30/30 [00:02<00:00, 14.39it/s]

from datasetsDynamic.loadDataBakery import loadDataBakery
data, XTrain, yTrain, XTest, yTest = loadDataBakery(testDays = 28, returnXY = True, daysToCut = 0, disable_progressbar = False)

Rolling: 100%|██████████| 152/152 [00:11<00:00, 13.25it/s]
Feature Extraction: 100%|██████████| 160/160 [00:43<00:00,  3.70it/s]
Rolling: 100%|██████████| 152/152 [00:12<00:00, 11.84it/s]
Feature Extraction: 100%|██████████| 160/160 [00:44<00:00,  3.59it/s]
Rolling: 100%|██████████| 152/152 [00:11<00:00, 13.53it/s]
Feature Extraction: 100%|██████████| 160/160 [00:44<00:00,  3.57it/s]