Machine learning has become more and more ordinary, and understanding which machine learning algorithm (or model type) to use, based upon your data and objectives, is important, and if you are relatively new to the process, it can be daunting.
Fitting a model to training data is one thing, but how do you know that the model (technique) or algorithm you select will generalize well to all your data and create the best prediction? Too much training or overfitting doesn't solve this problem; in fact, in this situation, it is typical for the model to perform poorly with totally new data.
Once again, the IBM Cloud platform provides robust and practical tools to assist you with this process.
The cloud offers a machine learning service (IBM Watson Machine Learning). The service offers the ability to manage your developed machine learning models using a continuous learning system, as well as an easy method for deployment, including online, batch, and streaming modes.
In addition, the model builder offered in IBM Watson Studio (which includes tutorials and sample datasets to illustrate how to train different types of machine learning models without the need for coding) can get you going quickly by stepping you through the, perhaps tedious, task of model selection (and even evaluation and deployment).