8/3/2019 Version Space Learning by Sammar Abbs

1/4

Report of

Version Space Learning

Submitted To:Mr. Saad Razzaq

Submitted By:

Sammar Abbas [29]

Arif Nawaz [28]

Iftikhar tarar [13]

Class:BSIT(Reg 5th)

Dept of CS & IT

University of Sargodha,Sargodha Pakistan

Date: January 19, 2012

8/3/2019 Version Space Learning by Sammar Abbs

2/4

1 Introduction

LearningLearning is the process that enables the system to do the same task more efficientlynext time. There are many types of learning,learning from being told,learning from

analogy,learning from discovery,and learning by examples.There are differenttechniques of learning by examples like we study decision tree. An other techniqueof learning by examples is version space.

Version space LearningVersion space is set of concepts consistent with a set of training examples is called

a version space (for that set of examples),and the version space learning is

hierarical representation of that knowledge/concepts got by sequence of learning

examples with out remembring any of the examples. The concept consistent must

include every positive instance, exclude every negative instance. The version space

method involves identifying all concepts consistent with a set of training example.

version space convergenceGeneralization and specialization leads to version space convergence.The key ideain version space learning is that specialization of the general models andgeneralization of the specific models may lead to just one correct model thatmatches all observed positive examples and does not match any negative examples.

2.Version space diagram

Top 1st

level: The top of tree, we have the most general hypothesis.

Top 2nd

level: This row is expanded form of first.This row of hypothsis is slightly more

specific then root nod.

Top 3rd

level: As training data (positive examples) is processed,the inconsistent nodes

are removed from general specification

Bottom 3rd

row: Any hypothesis taht is inconsistent with the training data (negative

example) is removed from tree.

Bottom 2nd

row: The specifice hypothesis is expanded to form more nodes that are

slightly more general.

Bottom 1st

row: This is most specific hypothesis.

VERSION SPACE DIAGRAM

8/3/2019 Version Space Learning by Sammar Abbs

3/4

3.Version space method

The version space method handles positive and negative examples symmetrically. We

have a representation language.A set of positive and negative examples expressed in

that language, and compute a concept description that is consistent with all the positive

examples and none of the negative examples.Accept a new training example.

If the example is positive,Generalize all the specific models to match the positive example,

but ensure the following.The new specific models involve minimal changes.Each new

specific model is a specialization of some general model.No new specific model is a

generalization of some other specific model.Prune away all the general models that fail to

match the positive example.

If the example is negative,Specialize all general models to prevent match with the negative

example, but ensure the followingThe new general models involve minimal changes.Eachnew general model is a generalization of some specific model.No new general model is a

specialization of some other general model.Prune away all the specific models that match

the negative example.

If S and G are identical, output their value and halt.if they are different, the training cases

were inconsistent.Output this result and halt.Else continue accepting new training

examples.

The algorithm stops when,It runs out of data.The number of hypotheses remaining is:

0 - no consistent description for the data in the language.

1 - answer (version space converges).

2+

- all descriptions in the language are implicitly included.

8/3/2019 Version Space Learning by Sammar Abbs

4/4

Exampels we solve by version space learning