Icon

JKI-S3-C18-FinnovationFlows

Parameter Optimization (Table) Component on a Generic Model with Range Sliders This component can optimize a generic classification model and a generic set of numerical parameters. In this case the component performs a parameter optimization on a gradient boosted to optimize number of trees ,max tree depth and learning rate. Check the workflow description to find a step-by-step guide to adapt the workflow to your own classification model.
Captures the Model training process
Using the best set of parameters for the model
Define Parameters Parameters to be optimized should be the same naming and number in those two nodes, names must start with GBT_ .
Parameter Optimization (Table)
Gradient Boosted Trees Predictor
Visualize
CSV Reader
0: train 1: test
Partitioning
Rule Engine
Table Transposer
Column Appender
Normalizer
Row Sampling
Column Filter
Gradient Boosted Trees Learner
LIME Loop Start
Capture Workflow End
Column Renamer
Gradient Boosted Trees Predictor
Bar Chart
Group Loop Start
Provide initial paramerter
Variable Creator
RowID
Capture Workflow Start
Loop End
Column Filter
Image to Table
conf matrix type
Rule Engine
Counter Generation
Read adult training data with class column.
CSV Reader
Supply parameters for optimization
Table Reader
training a local GLM for each input instance to generate a Local Inter. Model-agn. Explanation
Compute LIME
Color Manager
Loop End
Gradient Boosted Trees Predictor
Gradient Boosted Trees Learner
Select range of parameters (Click "Close & Apply" in the view after selection)
Parameter Ranges

Nodes

Extensions

Links