I have a dataset of 3,000 lines and 4,000 columns and the rows are grouped into 5 classes and I need to find out which columns, and combination of columns, are more discriminative for each combination of classes (re classifying into binary one versus the other or one combination versus all etc).
The best would be to first find the combination of features most relevant for a class (the identity of the features is what is of interest here), and then building a model to find and test the accuracy for predicting a class from those features.
I am looking for a machine learning expert in R so I can understand and replicate each step.
15 freelanceria on tarjonnut keskimäärin €88 tähän työhön
Dear client, I have a great experience in the field and would execute the task with a lot of efficiency. Kindly initiate a conversation and we could talk more about the details of your project. Best regards.