Machine Learning models are created to be configured and trained for performing specific tasks like making prediction or generating probability estimates.
Creating Machine Learning Models
After you prepare a dataset, you are ready to create a Machine Learning model. Start this by going to List Models page and click on the New button. This will take you to the model creation page.
Enter a name for the model, for example, Divorce Predictor; and also a description. Then specify the source tracker.
This will be a tracker you have already prepared to use as the dataset for training the model. See Preparing Machine Learning Dataset for information how to prepare a dataset tracker. Keep in mind that the selected source tracker can not be changed after the model has been created.
Using Model Template
Every model you create require further configuration that involves setting your preferred learner and transformers depending on the task the model is to perform. To make this process easier, Tiki includes some predefined configurations called model templates. To use a template, choose the one that matches your criteria from the Model template dropdown or select Start with a blank model to create the model without a template.
Tiki will create the model after you click on the Create button and will display the model configuration page. You can then go on to configuring the model as described in Configuring Machine Learning Models
Cloning an existing model
You can also create a new model from existing model through cloning. This approach is useful if you want to have a new model that is similar in configuration to an existing one. To clone a model, go to the List Models page and click on the Actions button of the model you wish to clone then select Clone.