comparison of Machine learning algorithms
Budget: £5 – £10 GBP
Realize and describe an experiment in R or Python that evaluates the classification performance for kNN and Naïve Bayes (see requirements above) on the Penguin dataset. Explain the choice (or necessity) of your error measurement method. Use appropriate illustrations and diagrams as well as statistics. For each of the considered models, perform the following steps:
1) Train the model with the train dataset.
2) Validate it with repeated 10-fold cross validation.
3) Tune hyperparameters (if applicable).
21
4) Test the model with the test dataset.
5) Compute the model performance.
2.2 Predictive performance of kNN algorithm using data processing
Realize and describe an experiment in R or Python that evaluates the classification performance of kNN using different data processing techniques:
• taking care of missing data (delete missing data or impute missing data)
• with PCA or without PCA
• with cross-validation (CV) or without CV (simple train-test split at random).
with dataset pamperpeguins
1) Train the model with the train dataset.
2) Validate it with repeated 10-fold cross validation.
3) Tune hyperparameters (if applicable).
21
4) Test the model with the test dataset.
5) Compute the model performance.
2.2 Predictive performance of kNN algorithm using data processing
Realize and describe an experiment in R or Python that evaluates the classification performance of kNN using different data processing techniques:
• taking care of missing data (delete missing data or impute missing data)
• with PCA or without PCA
• with cross-validation (CV) or without CV (simple train-test split at random).
with dataset pamperpeguins