SlideShare a Scribd company logo
Ir classification association
Automatic Classification

   Classification???
   Classificatory systems
   Output of such system
   Example of classification :
       Indexing
   Classification v/s Diagnosis ??
       Classification = grouping
       Diagnosis = identification
Classification Methods

   Classification Methods
       Why??
       Data
       Objects
           Documents , keywords, characters
       Data & objects
       Corresponding description
           attributes
Classification Methods
   Uses set of parameters to characterize each object
   Features should be relevant to task at hand
   Supervised classification
       What classes???
       Set of sample objects with known classes
   Training set
       Set of known objects
       Used by classification program
   Two phases for classification
       ??
       ??
Classification Methods

1.       Training Phase:
            Uses training set
            Decision is about
              How to weight parameters
              How to combine these objects under different classes


1.       Application Phase:
            Weights determined in phase 1 are used with set of objects
            That do not have known classes
            Determine their possible class
Classification Methods

   With few parameters ; process is easy
       Example:
   With much more parameters ; process is tough
       Example:
   Depending on structure ; find types of attributes
       Multi State Attribute
           Example:
       Binary State Attribute
        Example:
        
     Numerical Attributes
           Example
Classification Methods

   Binary State
       Bold , underline
   Multi State
       Color , position , font type
   Execution of operation changes attribute value.
   Example:
       MOVE
       FILL
       INSERT
       DELETE
       CREATE
Classification Methods
   Relation between Classes & Properties
    1.    Monothetic:
          To get membership of class ,
          object must posses the set of properties
          which are necessary as well as sufficient
          Example


    1.    Polythetic:
          Large number of members have some number of
           properties
          No individual is having all the properties
          example
Classification Methods

   Relation between Object & Classes
    1.    Exclusive:
          Object belongs to single class
          Example


    1.    Overlapping:
          Membership is with different classes
          Example
Classification Methods

   Relationship between Classes & Classes:
    1.    Ordered:
          Structure is imposed
          Hierarchical structure
          Example


    1.    Unordered:
          No imposed structure
          All are at same level
          example
Measures of Association

   Some classification methods are based on a binary
    relationship between objects

   On the basis of this relationship a classification method
    can construct a system of clusters

   Relationship type:
    1.   similarity
    2.   dissimilarity
    3.   association
Measures of Association

   Similarity:
       The measure of similarity is designed to quantify the likeness
        between objects
       so that if one assumes it is possible to group objects in such a
        way that an object in a group is more like the other members of
        the group
       than it is like any object outside the group,
       then a cluster method enables such a group structure to be
        discovered.
Measures of Association

   Association:
     Association means???
     Dependency…
     Occurrence…
     reserved for the similarity between objects
      characterized by discrete-state attributes.
Measures of Association

   Used to measure strength of relationship
   measure of association increases as the number or
    proportion of shared attribute states increases.
   Five measures of association
    1.   Simple
    2.   Dice’s coefficient
    3.   Saccard’s coefficient
    4.   Cosine coefficient
    5.   Overlap coefficient
Measures of Association

   Used in information and data retrieval
   | | specifies size of set
Probabilistic Indexing

   Probability of relevance
   Experiments and observations
   Sample space
   May Consist relevant as well as non relevant objects
   Consider a document
   Find no. of relevant document with respect to it
   That gives probability quotient
   probability measured as per the terms present in
    document
Probabilistic Indexing

   Probabilistic indexing model
   Contains random variable
   Denotes no. of relevant documents
   If this variable is selected by system
   Gives possible relevant document description
   Probabilistic information retrieval models are based on the
    probabilistic ranking principle,
   which says that documents should be ranked according to
    their probability of relevance with respect to the actual
    request.
Ir classification association
Ir classification association
Ir classification association
Ir classification association

More Related Content

PPTX
OPENPhacts target search - Mes & Friedeheim
PDF
It 405 materi 3 objek dan kelas
DOCX
วิเคราะห์แผน เรื่อง หน่วยการวัดความยาว
PPTX
sarah Place based project learning
PPTX
Everything Out Organizing Style Personality Preference
PDF
Sumários Desenvolvidos de Filosofia do Direito
DOC
PDF
Shopsial TVSS week 4
OPENPhacts target search - Mes & Friedeheim
It 405 materi 3 objek dan kelas
วิเคราะห์แผน เรื่อง หน่วยการวัดความยาว
sarah Place based project learning
Everything Out Organizing Style Personality Preference
Sumários Desenvolvidos de Filosofia do Direito
Shopsial TVSS week 4

Viewers also liked (12)

PDF
Requirements For Epcs When Marketing Homes For Sale Or Let
PPT
корпоративная культура
PPSX
Creative and Fun Photographs by John Wilhelm
PPSX
AS TAREFAS DO NADAL
PPT
Introducción al Hardware
PPT
來生,再也不愛你.Pps
PPT
Glori A
PDF
AoU CWEEN NEW ORLEANS Combined
PPTX
L'orto e il mare in barattolo
PDF
Waleed et al
PPTX
Descobrint picasso
DOC
Requirements For Epcs When Marketing Homes For Sale Or Let
корпоративная культура
Creative and Fun Photographs by John Wilhelm
AS TAREFAS DO NADAL
Introducción al Hardware
來生,再也不愛你.Pps
Glori A
AoU CWEEN NEW ORLEANS Combined
L'orto e il mare in barattolo
Waleed et al
Descobrint picasso
Ad

Similar to Ir classification association (20)

PPT
1 1 5 Clases
 
PPTX
CHAPTER 3 oop with programming java language
DOCX
Classification vs clustering
PPTX
SAP-ABAP-Object-Oriented-Programming.pptx
PPTX
Introduction to OOP with java
PDF
oops-123991513147-phpapp02.pdf
DOCX
Object oriented basics
PPTX
OOSD1-unit1_1_16_09.pptx
PPTX
System Concepts for Object Modelling.pptx
DOCX
Concept of Classification in Data Mining.docx
PPTX
Object oriented programming CLASSES-AND-OBJECTS.pptx
PDF
Java defining classes
PPT
automatic classification in information retrieval
PDF
Abap object-oriented-programming-tutorials
PPTX
Presentation
DOCX
Dbms question (3)
PPT
Object Oriented Design
PPT
Object Oriented Design
1 1 5 Clases
 
CHAPTER 3 oop with programming java language
Classification vs clustering
SAP-ABAP-Object-Oriented-Programming.pptx
Introduction to OOP with java
oops-123991513147-phpapp02.pdf
Object oriented basics
OOSD1-unit1_1_16_09.pptx
System Concepts for Object Modelling.pptx
Concept of Classification in Data Mining.docx
Object oriented programming CLASSES-AND-OBJECTS.pptx
Java defining classes
automatic classification in information retrieval
Abap object-oriented-programming-tutorials
Presentation
Dbms question (3)
Object Oriented Design
Object Oriented Design
Ad

More from swapnil shinde (8)

PDF
PDF
S.e 2003
PDF
Network analysis q.papers
PDF
Sw project mgmn q.papers
PDF
Daa q.paper
PDF
Ooad q.papers
PDF
Se oct2011
S.e 2003
Network analysis q.papers
Sw project mgmn q.papers
Daa q.paper
Ooad q.papers
Se oct2011

Recently uploaded (20)

PPTX
A powerpoint presentation on the Revised K-10 Science Shaping Paper
PDF
Complications of Minimal Access Surgery at WLH
PPTX
Introduction to Building Materials
PPTX
Cell Types and Its function , kingdom of life
PPTX
Lesson notes of climatology university.
PPTX
202450812 BayCHI UCSC-SV 20250812 v17.pptx
DOC
Soft-furnishing-By-Architect-A.F.M.Mohiuddin-Akhand.doc
PDF
Chinmaya Tiranga quiz Grand Finale.pdf
PDF
Classroom Observation Tools for Teachers
PPTX
CHAPTER IV. MAN AND BIOSPHERE AND ITS TOTALITY.pptx
PDF
Hazard Identification & Risk Assessment .pdf
PDF
LDMMIA Reiki Yoga Finals Review Spring Summer
PDF
Practical Manual AGRO-233 Principles and Practices of Natural Farming
PPTX
Final Presentation General Medicine 03-08-2024.pptx
PPTX
Final Presentation General Medicine 03-08-2024.pptx
PDF
A GUIDE TO GENETICS FOR UNDERGRADUATE MEDICAL STUDENTS
PPTX
Onco Emergencies - Spinal cord compression Superior vena cava syndrome Febr...
PDF
Weekly quiz Compilation Jan -July 25.pdf
PPTX
Unit 4 Skeletal System.ppt.pptxopresentatiom
PPTX
UV-Visible spectroscopy..pptx UV-Visible Spectroscopy – Electronic Transition...
A powerpoint presentation on the Revised K-10 Science Shaping Paper
Complications of Minimal Access Surgery at WLH
Introduction to Building Materials
Cell Types and Its function , kingdom of life
Lesson notes of climatology university.
202450812 BayCHI UCSC-SV 20250812 v17.pptx
Soft-furnishing-By-Architect-A.F.M.Mohiuddin-Akhand.doc
Chinmaya Tiranga quiz Grand Finale.pdf
Classroom Observation Tools for Teachers
CHAPTER IV. MAN AND BIOSPHERE AND ITS TOTALITY.pptx
Hazard Identification & Risk Assessment .pdf
LDMMIA Reiki Yoga Finals Review Spring Summer
Practical Manual AGRO-233 Principles and Practices of Natural Farming
Final Presentation General Medicine 03-08-2024.pptx
Final Presentation General Medicine 03-08-2024.pptx
A GUIDE TO GENETICS FOR UNDERGRADUATE MEDICAL STUDENTS
Onco Emergencies - Spinal cord compression Superior vena cava syndrome Febr...
Weekly quiz Compilation Jan -July 25.pdf
Unit 4 Skeletal System.ppt.pptxopresentatiom
UV-Visible spectroscopy..pptx UV-Visible Spectroscopy – Electronic Transition...

Ir classification association

  • 2. Automatic Classification  Classification???  Classificatory systems  Output of such system  Example of classification :  Indexing  Classification v/s Diagnosis ??  Classification = grouping  Diagnosis = identification
  • 3. Classification Methods  Classification Methods  Why??  Data  Objects  Documents , keywords, characters  Data & objects  Corresponding description  attributes
  • 4. Classification Methods  Uses set of parameters to characterize each object  Features should be relevant to task at hand  Supervised classification  What classes???  Set of sample objects with known classes  Training set  Set of known objects  Used by classification program  Two phases for classification  ??  ??
  • 5. Classification Methods 1. Training Phase:  Uses training set  Decision is about  How to weight parameters  How to combine these objects under different classes 1. Application Phase:  Weights determined in phase 1 are used with set of objects  That do not have known classes  Determine their possible class
  • 6. Classification Methods  With few parameters ; process is easy  Example:  With much more parameters ; process is tough  Example:  Depending on structure ; find types of attributes  Multi State Attribute  Example:  Binary State Attribute Example:   Numerical Attributes  Example
  • 7. Classification Methods  Binary State  Bold , underline  Multi State  Color , position , font type  Execution of operation changes attribute value.  Example:  MOVE  FILL  INSERT  DELETE  CREATE
  • 8. Classification Methods  Relation between Classes & Properties 1. Monothetic:  To get membership of class ,  object must posses the set of properties  which are necessary as well as sufficient  Example 1. Polythetic:  Large number of members have some number of properties  No individual is having all the properties  example
  • 9. Classification Methods  Relation between Object & Classes 1. Exclusive:  Object belongs to single class  Example 1. Overlapping:  Membership is with different classes  Example
  • 10. Classification Methods  Relationship between Classes & Classes: 1. Ordered:  Structure is imposed  Hierarchical structure  Example 1. Unordered:  No imposed structure  All are at same level  example
  • 11. Measures of Association  Some classification methods are based on a binary relationship between objects  On the basis of this relationship a classification method can construct a system of clusters  Relationship type: 1. similarity 2. dissimilarity 3. association
  • 12. Measures of Association  Similarity:  The measure of similarity is designed to quantify the likeness between objects  so that if one assumes it is possible to group objects in such a way that an object in a group is more like the other members of the group  than it is like any object outside the group,  then a cluster method enables such a group structure to be discovered.
  • 13. Measures of Association  Association:  Association means???  Dependency…  Occurrence…  reserved for the similarity between objects characterized by discrete-state attributes.
  • 14. Measures of Association  Used to measure strength of relationship  measure of association increases as the number or proportion of shared attribute states increases.  Five measures of association 1. Simple 2. Dice’s coefficient 3. Saccard’s coefficient 4. Cosine coefficient 5. Overlap coefficient
  • 15. Measures of Association  Used in information and data retrieval  | | specifies size of set
  • 16. Probabilistic Indexing  Probability of relevance  Experiments and observations  Sample space  May Consist relevant as well as non relevant objects  Consider a document  Find no. of relevant document with respect to it  That gives probability quotient  probability measured as per the terms present in document
  • 17. Probabilistic Indexing  Probabilistic indexing model  Contains random variable  Denotes no. of relevant documents  If this variable is selected by system  Gives possible relevant document description  Probabilistic information retrieval models are based on the probabilistic ranking principle,  which says that documents should be ranked according to their probability of relevance with respect to the actual request.