History: Training Machine Learning Models

            ^This page ((needs review))^

Machine Learning models require training before they can be used to perform tasks.

! Training Machine Learning Models
Training a model in Tiki is a one-step process. The model must have been created and fully configured with a data source tracker to be used for training. Tiki will check if the set source tracker contain any items (samples), and will halt the training process if there are none, and display an error message.

To train a model, go to the __List Models__ page and click on the __Actions__ button of the model, then select __Train__.

{img src="display1914" link="display1914" width="400" rel="box[g]" imalign="center" desc="Click on Train in model's action menu to start training" align="center" styleimage="border"}

Model will be trained using items from source tracker and a success message will be shown on completion.

{img src="display1915" link="display1915" width="400" rel="box[g]" imalign="center" desc="Success message after training" align="center" styleimage="border"}

!! Testing the Model
Testing a model in Tiki is the same as training except that only a sample of the model is created and it is not saved. Call it dry training.

{img src="display1916" link="display1916" width="400" rel="box[g]" imalign="center" desc="Clicking on Test in model's action menu will start test" align="center" styleimage="border"}

{img src="display1917" link="display1917" width="400" rel="box[g]" imalign="center" desc="Test complete" align="center" styleimage="border"}

This feature is useful in checking to know if model can learn with the linked source tracker. To test a model, open the model's action menu and click on __Test__.

!! Related links
* ((Machine Learning))
* ((Preparing Machine Learning Dataset))
* ((Creating Machine Learning Models))
* ((Configuring Machine Learning Models))
* ((Using Machine Learning Models))