Trained Model

A Trained Model is a software function produced by a model learning task/prediction task.

• AKA: Fitted Model.
• Context:
  • It can be assessed with a Model Evaluation Task.
  • It can be produced by a Model Training System (that solves a model training task).
  • . It can be serialized by an ML Model Serialization System (such as Mleap).
  • It can be managed by a ML Model Management System.
  • It can be deployed by an ML Model Deployment System.
  • It can be a component of a Trained Model Workflow (along with model data preparation code).
  • …
• Example(s):
  • a Trained Classification Model, such as the trained decision tree (trained on the iris dataset).
  • a Trained Ranking Function.
  • a Fitted Regression Model, such as the Regressed Linear Model of LLSE[math]\displaystyle{ (\lt 1,6\gt , \lt 2,5\gt , \lt 3,7\gt , \lt 4,10\gt ) \Rightarrow y=1.4x+3.5 }[/math]
  • a Trained Clustering Model, such as a trained k-means model.
  • a Fitted Latent Factors Model.
  • a Pre-Trained Model.
  • …
• Counter-Example(s):
  • an Untrained Model.
  • a Machine Learning Model Family, such as a Decision Tree Metamodel.
• See: Learned Predictor, Learned Data Model, Ontology Learning Task, Metamodel Design Task, Modeling Language, Statistical Model, Deductive Model.

---

---

References

2011

• http://code.google.com/apis/predict/docs/glossary.html
  • When you train the Prediction API against a data set, it creates a model based on that specific training data. All queries for the Prediction API are sent to a model trained against a specific training data.

---