ANN - Learning rateIn the discussion among peers about an ANN analysis, the questions arose:
- What learning rate was used in the analysis? Since I didn't use boosting (number of models is 0), is the 0.1 learning rate predefined by the platform relevant?
- What means the value for the sum of squares to stop the calculation. Is it the sum of squares error (SSE) that is reported for each response?
Thank you so much!Advanced Statistical ModelingRe: ANN - Learning rateIf you didn't use boosting then the learning rate is not used and has no impact on the analysis.
The reported SSE is the for the sum of the squared differences of the actual response and the predicted response, for each row of data used in the analysis.
