Skip to content

ML Model#

🔒 MLModel.Show

Overview#

The ML Model (short for Machine Learning Model) entity represents a machine learning model. In addition to the trained model, this object contains all the settings needed to transform a given Data Set (represented in the system through the Data Set entity) into data that can be used to train/validate/test a model using AutoML. The model itself can be deployed so that it can output predictions in real time.

Warning

ML Model depends on a containerized installation and is not available for traditional installations.

Tying Everything Together#

graph LR
    A1[Data Set] --- Main[ML Model]

classDef mermaid_title color:#000, fill:#fafafa, stroke:#fafafa, stroke-width:0x, font-size:100%, font-weight:200;
classDef mermaid_start color:#000, fill:#fafafa, stroke:#fafafa, color:#fafafa, stroke-width:0x, font-size:100%, visibility: hidden;
classDef mermaid_businessdata color:#000, fill:#65CDE8, stroke:#65CDE8, stroke-width:0px, font-size:100%;
classDef mermaid_nonbusinessdata color:#000, fill:#B7DEE8, stroke:#B7DEE8, stroke-width:0px, font-size:100%;
classDef mermaid_entity color:#000, fill:#FB9F53, stroke:#FB9F53, stroke-width:0px, font-size:100%;
classDef mermaid_entitylinked color:#000, fill:#FCD5B5, stroke:#FCD5B5, stroke-width:0px, font-size:100%;
classDef mermaid_context color:#000, fill:#B9CDE5, stroke:#B9CDE5, stroke-width:0px, font-size:100%;
classDef mermaid_optional color:#000, fill:#B7DEE8, stroke:#65CDE8, stroke-width:1px, font-size:100%, stroke-dasharray: 5 5;
class Main mermaid_entity
class A1,A2,A3,A4,A5,A6,A7,A8,A9,A10 mermaid_businessdata
class L1,L2,L3,L4,L5,L6 mermaid_entitylinked
class C1,C2,C3,C4,C5,C6 mermaid_context
class N1,N2,N3,N4,N5,N6 mermaid_nonbusinessdata

click Main "../../business-data/ml-model"
click A1 "../../business-data/data-set"

Sequence Of Steps#

The necessary steps to create an ML Model are:

  1. Pick a Change Set
  2. Name the ML Model
  3. Select the Data Set that should be used to train the model
  4. Select the Data Set fields needed to train the model, and for each field select the transformations that should be applied to the data
  5. Pick a field to predict by marking it as Label
  6. Select the percentage of data that should be used to train/validate/test the model
  7. Transform the data
  8. Train the model by picking how much time the AutoML algorithm should run and the metric to optimize; AutoML will automatically train numerous models using different machine learning algorithms with varying hyperparameters, and then return the best model according to the specified metric
  9. Once trained, the model can be tested to evaluate its accuracy
  10. Finally, the model can be deployed so that it can start outputting predictions in real time