featurey.py model.py get_product() tr = build_model() build_data() build_x() build_y() predict(tr, te) score()
output format (the same day) { "stock id 1": { "feature1": 0, "feature2": 0, "feature3": 0, "feature4": 0, "feature5": 0, }, "stock id 2": { "feature1": 0, "feature2": 0, "feature3": 0, "feature4": 0, "feature5": 0, }, }
build_x input:
- start day: range of sampling
- end day: range of sampling
- seleced features: type = list, element = 'feature1', 'feature3'
- features days: days per sample
output:
- X: list, [2, 30, 3, ... ], features per day * features day