Hi Xun,
The OnlineLogisticRegression model data is represented as Flink Tables. A Flink
table can be unbounded and contains continuous model data, so the way to update
model is just insert new model data into the Table. Actually,
OnlineLogisticRegression can fit models continuously from unbounded input data,
so it is recommended to use OnlineLogisticRegression to generate a Table of
continuous models. The OnlineLogisticRegressionExample is an example of using
OnlineLogisticRegression on bounded input data.
