1) Do I need to split the data to training set and test set? 2) Use the actual results to test the accuracy of the algorithm?

Question

I am new to machine learning. I have a dataset where I know the output of the data (label dataset). I have to develop an unsupervised learning model and more specific Isolation