Using Gaussian Processes to predict XOR data

You can easily download and run the workflow directly in your KNIME installation. We recommend that you use the latest version of the KNIME Analytics Platform for optimal performance.

Here's how the workflow operates:

1. Python Script node generates a dataset with 200 random points in a 2D space. Target variable "Y" is then generated based on the XOR logic function. "Y_nominal" is the nominal form ("yes" or "no") of the target variable "Y", and "Y_numeric" is in the corresponding numeric form (1 or 0).

2. Then we split the dataset into train and test subsets.

3. Lasso Regression is performed with feature targets X_0, X_1, and target column Y_numeric.

Gaussian Process Regression is performed with feature targets X_0, X_1, and target column Y_numeric.

Gaussian Process Classification is then performed with feature targets X_0, X_1, and target column Y_nominal.

4. For each algorithm, a Python view is created showcasing a plot with data points coloured based on their class.

External resources

• Scikit-Learn - Lasso Regression
• Scikit-Learn - Gaussian Process Regression
• Scikit-Learn - Gaussian Process Classification
• Scikit-Learn - GPC on the XOR dataset