From the course: Certified Analytics Professional (CAP) Cert Prep

Unlock the full course today

Join today to access over 24,600 courses taught by industry experts.

Recalibrating the model through validation

Recalibrating the model through validation

From the course: Certified Analytics Professional (CAP) Cert Prep

Start my 1-month free trial Buy for my team

Recalibrating the model through validation

“

- [Instructor] Recalibrating analytics models is a necessary and natural step after tracking them for potential deterioration. Validation is a crucial part of recalibration because it allows us to measure performance improvements after adjusting a model for better outcomes. In predictive analytics, we divide data into training and test sets. As its name suggests, the training data set is for training a machine learning or ML algorithm. While the test data sets are for validating the performance of the ML model in use. We call this approach holdout because the subset of the data is held out for testing. There is a trade off in this method. The more data you use for training, the less data left for testing. Also, there is only a single opportunity to test the model. Cross-validation is a good way to overcome these limitations. It partitions a dataset into K segments also called folds. K here is a placeholder for a natural…

Contents