Model Training

Overview

Model Training defines how QMI predictions are generated for each evaluation period. Once per month, you load historical data into CRM Analytics, train the model, and calculate QMI scores for the next period. The goal is to predict and maximize the Quality Metric Index (QMI).

During training, you control how the model is built and evaluated. You can select metrics, choose a training algorithm, or run a model tournament to compare results. The data splits into training and validation sets, and the model reports accuracy metrics such as R-squared, which depend on data quality.