scoring-model

This model automates confidence scoring for stock recommendations by processing input evidence and applying proprietary scoring rules to generate a confidence score and detailed breakdown. It uses an additive-with-saturating-caps model.

The additive-with-saturating-caps model is a proprietary scoring method used to compute confidence scores for stock recommendations. It processes input evidence to generate both a final confidence score and a detailed breakdown.

Yes, you can adjust the confidence scoring model using a manual weight-refit procedure. This allows you to handle various stock scenarios and tailor the scoring rules to your specific investment data.

Yes, computing confidence scores requires access to specific configuration files and input data. The model depends on data/config.json to process evidence and apply the necessary scoring rules.

You should trigger the confidence scoring model during Phase C or Phase E of the india-stock-recommender pipeline, or when a user requests an explanation of confidence scores or an audit of score breakdowns.