Think of online and batch machine learning concepts as basically the difference between performing multiple iterations of updating predictor values from new chunks of data compared to churning through all of the available data first, and then setting the predictor values:

• Online machine learning: This is a technique of machine learning where data are made available in sequential order and is used to streamline the best predictor for future data at each step or iteration.
• Batch learning: Batch machine learning is a method that will generate the best predictor by learning on the entire training dataset at once.