I am using the Smile library in Scala (similar story holds in Java) to compute a linear regression model of some training examples. I create the model using the ridge function.
val x = readCsv._1 // contains the features
val y = readCsv._2 // contains the label
val model = ridge(x, y, 0.0057)
My question: is it possible to retrain such a model using new labeled training examples or do the old ones have to be included in the retraining as well?