AdaBoost Model

Command:

Math -> AdaBoost Model...

The command Math/AdaBoost Model... allows calculating an AdaBoost model (both as a "regression" model and as a classification model, depending on the target variable). Please note that the AdaBoost model is only available if you have Python installed on your system.

In order to calculate a model you first have have to specify the variables by clicking into the corresponding variable list. The independent variables (descriptors) are at the left, the dependent variable (target variable) can be selected at the right. After clicking the variable list the variable selection dialog allows to select the desired variables.

In order to calculate the Adaboost model the button "Calculate" () has to be pressed. After that the most important diagrams for checking the results are displayed in the switchable diagnostic window:

• Actual vs. Estimated: This tab shows the estimated target data plotted against the actual ones.
• Residuals: The corresponding residuals.
• Distribution of Residuals:
• Tree Scan: The Tree Scan tab allows to find the optimum number of trees. Click "Start Tree Scan" to scan the number of trees. Please note that the Tree Scan results depend on the learning rate (LR) parameter. Thus you have to repeat the scan when you change the LR parameter.
• LRate Scan: The LRate Scan tab allows to find the best learning rate. Click "Start LRate Scan" to scan the entire allowed range of learning rate values. Please note that the LRate Scan results depend on the number of trees. Thus you have to repeat the scan when you change the number of trees.

How To:

Training an AdaBoost model is straightforward, please follow these steps: Select the descriptors and the target variable. The target variable has to be continuous. Set the number of trees, the tree depth and the learning rate. In most the cases, the default values will work well. Click "Calculate" to start the calculation of the AdaBoost model.