Data Collection

Overview

An Einstein predictive model requires a dataset with a minimum of 400 rows to support reliable prediction. This requirement ensures the model has enough historical data to identify patterns.

Before model creation, the data undergoes no cleaning other than filtering records based on void conditions. You must ensure the dataset quality and completeness before using it for prediction.

You must create a Data Model before you can view the Prediction Score.

For setup instructions, see QMI Settings and follow the steps to create the required Data Model.