which part of "train_model" trains a meta model?

[dualwieldingrogue]

Based on the description in: backtester/training.py:

"Trains a meta-model using labeled financial events and returns structured model output."

But based on the code, and how it's used in "Part_3-Labeling", there's no primary model input as part of the feature anywhere.

The "get_trend_scanning_meta_labels" function appears to provide meta labels, but that function is somehow commented off in the notebook, and no longer exists in the code.

[pnjoroge54]

Hi, I removed it because I thought better of what creating meta-labels from trend-scanning labels means. Their aim is too have a model that let's trends run until they're exhausted. As you noticed, in the article I used the labels as a filter. I will organize this code base better soon. For to long I was making changes to the main branch.