Question

I’ve implemented a classifier which

• Each iteration receives a parameter object to classify, some objects share a classifiable “property” like a color name.
• Classification parameters could change, so they are parametrized too
  and passed to this classifier at initialization time.
• The classifier implements the classification logic based in the type
  of possible classifications AND the object to classify.

I am VERY confused about the vocabulary used in some articles: Linear Classifier, Feature Values and Vectors.

Is this a common form of classifier?
In my approach I see no vectors, no statistical classifications, no hierarchical classificatoin, no machine learning, etc.

Which kind of classifier would be for the computer science literature?

Answer 1

Your “parameter object” is a feature vector. Your classifier apparently does not involve training so I’d say it’s an ad hoc rule-based classifier.