How to interpret the graph in model simulator

k_vishnu772

HI All, I ran the model simulator and got the result but there is a graph showing support pridiction and contadicts prediction (i attached here) what are those values in the graph represents is it percentage ? it does not 100 % when i sum them up,COuld you please help me

IngoRM

Hi,

 

Check out this documentation page here:

 

https://docs.rapidminer.com/latest/studio/operators/scoring/explain_predictions.html

 

The answer is a bit hidden in this statement here: "For each Example in an ExampleSet, this operator generates a neigboring set of data points, and uses correlation to identify the local attribute weights in that neighborhood."

 

So these are correlations of each attribute to the predictions in the neighborhood of the data point. Here are some more details on this:

 

This algorithm creates a local explanation of why a prediction was made for each example of a given input set. The algorithm uses a local sample of data points around each data point which should be explained. We create predictions of those neighbors with the model.  Then we calculate the local correlation of all attributes to those predictions to identify the local attribute weights in this region. Those correlations are then translated to the green and red bars.

 

Best,

Ingo

IngoRM

Hi,

 

Check out this documentation page here:

 

https://docs.rapidminer.com/latest/studio/operators/scoring/explain_predictions.html

 

The answer is a bit hidden in this statement here: "For each Example in an ExampleSet, this operator generates a neigboring set of data points, and uses correlation to identify the local attribute weights in that neighborhood."

 

So these are correlations of each attribute to the predictions in the neighborhood of the data point. Here are some more details on this:

 

This algorithm creates a local explanation of why a prediction was made for each example of a given input set. The algorithm uses a local sample of data points around each data point which should be explained. We create predictions of those neighbors with the model.  Then we calculate the local correlation of all attributes to those predictions to identify the local attribute weights in this region. Those correlations are then translated to the green and red bars.

 

Best,

Ingo