Model Monitor
Model Monitoring refers to the process of closely tracking the performance of
models in production and helps you to frequently monitor the distribution of the model
and provides alerts in case of any exceptions.
In addition, it enables you to identify and eliminate bad quality
predictions and poor technical performance of the models.
The model's robustness depends not only on the training of the
feature-engineered data but also on how well the model is monitored after deployment.
Typically a model's performance degrades over time, and it essential to detect the cause
of the decrease in performance of the model. The main cause of the decline in
performance can be drift in the independent or/and dependent features which may violate
the model’s assumption and distribution about the data. When models are built, the model
builder will create a snapshot of the dataset used for training the model and save it in
a file which is later used for calculating drift with the current snapshot of the
dataset.
To start the model monitoring, perform the following steps: