Instance vs Model based learning

 instance model:

   the point to be predicted or to be analyzed checks its neighbor point and compares its probability or output value: If the output value is success or positive in case of compared value than result of that point will be also success else fails

model based:

 it creates certain pattern and if the predicted value falls under the desired pattern model gives success result else not success