A Look at AutoML Models in the Platform

Goal

With this functionality, you will be able to select the attributes of an Entity (or make an SQL query on one or several entities) and create and train a learning model, which will allow you to predict another attribute.

Besides, these models can be APIfied through the API Manager

How it works on the Platform

There is a new feature, “My AutoML Predictors”:

in which you can see my Predictors and the state in which they are:

You can also create a Predictor on an Entity, selecting the input attributes to the model of an ontology or a SQL query on one or several Entities, and selecting the field to be predicted:

you can see how the model is being trained:

and at the end you will see the selected model among the tested ones and its accuracy:

y al concluir veré el modelo seleccionado entre los probados y su accuracy:

If you click on the detail, you can see more information:

And you will now be able to query the predictions, for example from the QueryTool:

SELECT * FROM iris_predictor2 where when_data='{"sepal_length":5.1, "sepal_width":3.5, "petal_length": 1.4, "petal_width":0.2}'