The Model Builder in IBM Watson Studio is a graphical tool that actually guides you, step by step, through building your first machine learning models. The model builder utilizes the following workflow:
- Upload data to train
- Choose a machine learning technique and algorithm
- Train and evaluate the model
- Test and deploy the model
The model builder (currently) focuses on creating three basic types of machine learning model techniques (which is typically more than sufficient for starting out with most machine learning projects). Furthermore, for each kind of model, you can choose from multiple algorithms to implement within the model. These are referred to as model techniques.
Part of the angst of building a machine learning solution is selecting the proper ML algorithm to be used. If unsure, or in the interest of saving time, at least for your first few attempts, you may want to utilize the option of having the model builder automatically select an algorithm for you, based upon the training data you provide. The model techniques include the following:
- Binary classifier: Classifies data into two categories
- Multiclass classifier: Classifies data into multiple categories
- Regression: Predicts a value from a continuous set of values