Articles on: Basic concepts

5. Batch Learning versus Incremental Learning

We can classify machine learning models by whether the system learns the data incrementally from a flow of live data.

Batch Learning

Batch machine learning refers to the process of training a model on a dataset that is fully available at the beginning of the learning process. The model is then deployed and used without any further updates. The main advantage of batch learning is that it can be done using more powerful hardware and can be optimized for specific use cases. Additionally, the training dataset can be thoroughly cleaned and preprocessed, leading to higher accuracy of the model.

Incremental (Online) Learning

Incremental machine learning, on the other hand, refers to the process of training a model on data as it is being collected, rather than using all the data at once. This allows the model to adapt to new data and improve over time. One of the main advantages of incremental learning is that it can handle streaming data, and can learn from very large datasets. It also allows for the model to be updated in real-time with new data, which is useful in applications such as online advertising or recommendation systems.

Both online and offline machine learning have their advantages and disadvantages. Online learning is more adaptable and can handle changing data, but it can also be more prone to overfitting and require more computational resources. Offline learning can be optimized for specific use cases and can lead to higher accuracy, but it doesn't handle changing data as well, and the model may become outdated over time.

Updated on: 25/01/2023

Was this article helpful?

Share your feedback


Thank you!