Next: Approaches to classification
Up: Cont'd
Previous: Data Mining operations
- Supervised Learning,inductive learning
Inductive learning is the technique to infer information that
is generalised from the database.It is a higher level information
in that is a general statement about objects in the database
- Examples are described in terms of attributes(also called features,
variables,dimensions). An attribute's range can be any one of:
- Categorical (range is unordered set of symbolic values,possibly
partially ordered)
- Ordinal (range is linearly totally ordered set of symbolic values
- Numeric or continuous(a dense subset of the reals)
- Class(output variable):categorical values.If the class variable is
numeric,then the problem becomes a regression problem.This is an
important distinction as we shall see later.
- Input is in form of feature vectors.If problem has k features then
training data consists of k-feature vectors.A class is given for
each example (class vector of data set)
DBMS
1999-03-24