Load dataset diabetes.arff, to be found in the "Data" folder in WEKA (download folder), also loaded in the folder "Assignments&Labs".
others
Description
Load dataset diabetes.arff, to be found in the "Data" folder in WEKA (download folder), also
loaded in the folder "Assignments&Labs". For an explanation of the dataset see
https://archive.ics.uci.edu/ml/support/diabetes.
Apply Feature Selection using Principal Components Analysis (PCA).
Under Select Attributes -> Attribute Evaluator (select PrincipalComponents). Note the
default Search Method should be Ranker (so click yes at the pop-up or change the method
before selecting attribute evaluator. Leave all default settings and select Start. What are the top 3
features or feature-sets?
. Repeat the PCA, adjust the varianceCovered to 0.7, and run the model. Describe your findings.
Using the same data, Apply Normalization. Under Preprocess -> Filter-> Unsupervised->
Attribute->Normalize->Ok. Close and Then Apply. Use default settings. Then run PCA again
with 0.95 for varianceCovered. Explain your results.
Additionally, click ‘Save’ and save the normalized data as csv OR select ‘Edit’ and take a screenshot of
your normalized data.
Using the same data. Apply Binning. Under Preprocess -> Filter-> Unsupervised-> Attribute-
>Discretize ->Ok. Close and Then Apply. Use default settings. Then run PCA again with 0.95
for varianceCovered. Explain your findings. Note: Make sure to Undo the normalization
before you apply binning.
