Design a Radial Basis Kernel-Support Vector Machine (Kernel-SVM) based classification model for facial expression recognition tasks. Develop the model considering the following steps:

Data pre-processing:

Encode the attributes and

Demonstrates descriptive statistics of attributes

Model implementation: Train the Kernel-SVM method with training sets.

Evaluation: Test model with independent test sets (20% samples).

Write a report including training logs (if available), testing accuracy with different performance metrics (confusion matrix, precision, recall, F1-score, ROC curve, etc.) for evaluating classifier, and computational times for training and testing phases.

After vectoring the training and test sets, apply Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) and represent the features in the lower dimensional space (k <<<< 15). Then apply the SVM on both lower dimensional feature space and analysis the results.

Steps need to be considered:

Data pre-processing: apply the PCA and LDA to reduce the feature dimensionality.

Model implementation: train the Kernel-SVM method with new features generated using PCA and LDA

Testing sets preparation: apply the PCA and LDA on testing samples.

Evaluation: testing the models with independent test sets (20% samples).

Finally, compare the results generated from both methods and report accordingly. Performance metrics: confusion matrix, precision, recall, F1-score, ROC curve, etc.

Note: If you implement the PCA and LDA from scratch instead of sklearn, you will be given some bonus points for the second task.

Can you explain these with python code.