A naive predictor (KNN classifier) based on demo data for cancer diagnosis.
The KNN classifier / predictor is written in a simplified user friendly entrance (a python class). Users can choose from 3 distance algorithms (Euclidian, Mahattan or Chebyshev) and the number of neighbours (value of k), when creating the KNN object. The training process for the predictor initialization is done automatically by the KNN classifier based on the give data set (in "csv" file).
The graphic is based on the best prediction accuracy by different distance algorithms and k values.
