SlideShare una empresa de Scribd logo
1 de 12
Descargar para leer sin conexión
4/30/2012




Linear Discriminant Analysis
 Proposed by Fisher (1936) for
 classifying an observation into one of
 two possible groups based on many
 measurements x1,x2,…xp.

 Seek a linear transformation of the
  variables Y=w1x1+w2x2+..+wpxp + a constant




                                                      1
4/30/2012




  Linear Discriminant Analysis
  Discriminant analysis – creates an
   equation which will minimize the
   possibility of misclassifying cases into their
   respective groups or categories.




The purposes of discriminant analysis (DA)
   Discriminant Function Analysis (DA)
    undertakes the same task as multiple
    linear regression by predicting an
    outcome.
   However, multiple linear regression is
    limited to cases where the dependent is
    numerical
   But many interesting variables are
    categorical,



                                                           2
4/30/2012




  The objective of DA is to perform
   dimensionality reduction while preserving as
   much of the class discriminatory information
   as possible
  Assume we have a set of D-dimensional
   samples {x 1, x2, …, xN}, N1 of which belong to
   class ω1, and N2 to class ω2.
  We seek to obtain a scalar y by projecting
   the samples x onto a line
                    y = wTx




•The top two distributions overlap too much and do not
discriminate too well compared to the bottom set.
•Misclassification will be minimal in the lower pair,
•whereas many will be misclassified in the top pair.




                                                                3
4/30/2012




 Linear Discriminant Analysis
      Assume variance matrices equal
      Classify the item x at hand to one of J groups
      based on measurements on p predictors.

      Rule: Assign x to group j that has the
      closest mean
            j = 1, 2, …, J
      Distance Measure: Mahalanobis Distance.




  Linear Discriminant Analysis
  Distance Measure:
         For j = 1, 2, …, J, compute


         d j  x    x  x j Spl  x  x j
                             T    1


Assign x to the group for which dj is minimum
S is the pooled estimate of the covariance
 pl

      matrix




                                                              4
4/30/2012




…or equivalently, assign x to the
group for which


    L x   x S
                        1      1            1
                           x
                    T                    T
       j        j       pl
                                2   x S x
                                     j       pl   j



 is a maximum.
 (Notice the linear form of the equation!)




Linear Discriminant Analysis
…optimal if….
• Multivariate normal distribution for the
  observation in each of the groups
• Equal covariance matrix for all groups
• Equal prior probability for each group
• Equal costs for misclassification




                                                             5
4/30/2012




Relaxing the assumption of equal prior
probabilities…



L x   ln p j  x S
                             T           1
                                 1                       1
                                    x
                                                  T
  j
                         j
                                 pl
                                         2   x S x
                                              j           pl   j




 pj   being the prior probability for the jth
      group.




Relaxing the assumption of equal
covariance matrices…

                    1
  Q  x   ln p j  ln S
      j                                               j
                    2
   x  x  S x  x j 
                 T   1
             j       j




result?…Quadratic Discriminant
Analysis




                                                                          6
4/30/2012




Quadratic Discriminant Analysis

    Rule: assign to group j if Q x 
                                  j
                                         is
    the largest.

Optimal if
the J groups of measurements are
multivariate normal




 Other Extensions & Related Methods
  Relaxing the assumption of normality…
       Kernel density based LDA and QDA



  Other extensions…..
       Regularized discriminant analysis
       Penalized discriminant analysis
       Flexible discriminant analysis




                                                     7
4/30/2012




Evaluations of the Methods

     Classification Table (confusion matrix)

                               Predicted group
Actual group   Number of
               observations

                               A                 B



A              nA              n11               n12
B              nB              n21               n22




Evaluations of the Methods
Apparent Error Rate (APER):
                          # misclassified
               APER =
                          Total # of cases
….underestimates the actual error rate.

Improved estimate of APER:
Holdout Method or cross validation




                                                              8
4/30/2012




Fisher's iris dataset
  •The data were collected by Anderson and used
  by Fisher to formulate the linear discriminant
  analysis (LDA or DA).
  •The dataset gives the measurements in
  centimeters of the following variables:
  1- sepal length, 2- sepal width, 3- petal length,
  and 4- petal width,
  this for 50 fowers from each of the 3 species of
  iris considered.
  •The species considered are Iris setosa,
  versicolor, and virginica




 setosa          versicolor    virginica




                                                             9
4/30/2012




An Example: Fisher’s Iris Data

Actual               Number of             Predicted Group
                     Observations
Group
                                           Setosa        Versicolo   Virginica
                                                         r
Setosa               50                    50            0           0
Versicolor           50                    0             48          2
Virginica            50                    0             1           49

Table 1: Linear Discriminant Analysis
                  (APER = 0.0200)




  An Example: Fisher’s Iris Data


Actual               Number of             Predicted Group
                     Observations
Group
                                           Setosa        Versicolo   Virginica
                                                         r
Setosa               50                    50            0           0
Versicolor           50                    0             47          3
Virginica            50                    0             1           49

Table 1: Quadratic Discriminant Analysis
                 (APER = 0.0267)




                                                                                       10
4/30/2012




                   An Example: Fisher’s Iris Data
                   2.5
                                                                                v            v
                                                        v             v             v
                                                v               v     v    v        v
                                                        v       v                                    v
                                                        v       v     v         v
                   2.0




                                           v            v       v          v                         v
                                           v        v   v
                                           v        v   v   v   v     v    c
                                                                           v
                                           v                    c
                                                    v
                                                    c           v               c   c
                   1.5




                               c
                               v           c            c
                                                        v   c   c     c    c
                                                v   c   c   c   c     c    c
Petal Width




                                   c       c        c   c   c   c
                                                c   c   c       c
                                       c   c
                   1.0




                          c    c   c   c        c   c



                                                                                        s
                   0.5




                                                                                s
                                                                                    s            s   s   s                s
                                   s                            s                   s   s            s
                                                            s   s     s    s    s   s   s    s   s   s       s        s
                                                                s     s                      s                    s

                          2.
                           0               2.
                                            5                   3.
                                                                 0                      3.
                                                                                         5                   4.
                                                                                                              0
                                                                     S Wi t
                                                                      epal dh




              An Example: Fisher’s Iris Data
                    2.5




                                                                                o
                                                                                v            o
                                                                                             v
                                                        o
                                                        v              o
                                                                       v            o
                                                                                    v
                                                o
                                                v                o
                                                                 v     o
                                                                       v   o
                                                                           v        o
                                                                                    v
                                                        o
                                                        v        o
                                                                 v                                   o
                                                                                                     v
                                                        o
                                                        v        o
                                                                 v     o
                                                                       v        o
                                                                                v
                    2.0




                                            o
                                            v           o
                                                        v        o
                                                                 v         o
                                                                           v                         o
                                                                                                     v
                                            o
                                            v       o
                                                    v   o
                                                        v
                                            o
                                            v       o
                                                    v   o
                                                        v   o
                                                            v    o
                                                                 v     o
                                                                       v   o
                                                                           c
                                                                           v
                                            o
                                            v                    x
                                                                 c
                                                    o
                                                    c            o
                                                                 v              x
                                                                                c   x
                                                                                    c
                    1.5




                               x
                               o
                               c
                               v            x
                                            c           x
                                                        c
                                                        v   x
                                                            c    x
                                                                 c     x
                                                                       c   x
                                                                           c
                                                o
                                                v   x
                                                    c   x
                                                        c   x
                                                            c    x
                                                                 c     x
                                                                       c   x
                                                                           c
     Petal Width




                                   x
                                   c        x
                                            c       x
                                                    c   x
                                                        c   x
                                                            c    x
                                                                 c
                                                x
                                                c   x
                                                    c   x
                                                        c        x
                                                                 c
                                       x
                                       c    x
                                            c
                    1.0




                           x
                           c   x
                               c   x
                                   c   x
                                       c        x
                                                c   x
                                                    c



                                                                                         +
                                                                                         s
                    0.5




                                                                                +
                                                                                s
                                                                                    +
                                                                                    s            +
                                                                                                 s   +
                                                                                                     s   +
                                                                                                         s                +
                                                                                                                          s
                                   +
                                   s                             +
                                                                 s                  +
                                                                                    s    +
                                                                                         s           +
                                                                                                     s
                                                            +
                                                            s    +
                                                                 s     +
                                                                       s   +
                                                                           s    +
                                                                                s   +
                                                                                    s    +
                                                                                         s   +
                                                                                             s   +
                                                                                                 s   +
                                                                                                     s        +
                                                                                                              s       +
                                                                                                                      s
                                                                 +
                                                                 s     +
                                                                       s                     +
                                                                                             s                    +
                                                                                                                  s

                          2.
                           0               2.
                                            5                   3.
                                                                 0                      3.
                                                                                         5                   4.
                                                                                                              0
                                                                     S Wi t
                                                                      epal dh




                                                                                                                                    11
4/30/2012




Summary
LDA is a powerful tool available for
classification.
     Widely implemented through various
     software
     Theoretical properties well
  researched




                                                12

Más contenido relacionado

La actualidad más candente

Logistic regression
Logistic regressionLogistic regression
Logistic regressionDrZahid Khan
 
Principal Component Analysis and Clustering
Principal Component Analysis and ClusteringPrincipal Component Analysis and Clustering
Principal Component Analysis and ClusteringUsha Vijay
 
Discriminant analysis
Discriminant analysisDiscriminant analysis
Discriminant analysisWansuklangk
 
Regression analysis in R
Regression analysis in RRegression analysis in R
Regression analysis in RAlichy Sowmya
 
Introduction to Machine Learning Classifiers
Introduction to Machine Learning ClassifiersIntroduction to Machine Learning Classifiers
Introduction to Machine Learning ClassifiersFunctional Imperative
 
K mean-clustering algorithm
K mean-clustering algorithmK mean-clustering algorithm
K mean-clustering algorithmparry prabhu
 
Fishers linear discriminant for dimensionality reduction.
Fishers linear discriminant for dimensionality reduction.Fishers linear discriminant for dimensionality reduction.
Fishers linear discriminant for dimensionality reduction.Nurul Amin Choudhury
 
Analysis of variance ppt @ bec doms
Analysis of variance ppt @ bec domsAnalysis of variance ppt @ bec doms
Analysis of variance ppt @ bec domsBabasab Patil
 
Correlation and regression
Correlation and regressionCorrelation and regression
Correlation and regressionAjendra7846
 
Discriminant analysis group no. 4
Discriminant analysis  group no. 4Discriminant analysis  group no. 4
Discriminant analysis group no. 4Advait Bhobe
 
Introduction to Linear Discriminant Analysis
Introduction to Linear Discriminant AnalysisIntroduction to Linear Discriminant Analysis
Introduction to Linear Discriminant AnalysisJaclyn Kokx
 
Logistic regression
Logistic regressionLogistic regression
Logistic regressionsaba khan
 
Lect4 principal component analysis-I
Lect4 principal component analysis-ILect4 principal component analysis-I
Lect4 principal component analysis-Ihktripathy
 

La actualidad más candente (20)

Logistic regression
Logistic regressionLogistic regression
Logistic regression
 
Cluster analysis
Cluster analysisCluster analysis
Cluster analysis
 
Principal Component Analysis
Principal Component AnalysisPrincipal Component Analysis
Principal Component Analysis
 
Discriminant analysis
Discriminant analysisDiscriminant analysis
Discriminant analysis
 
Principal Component Analysis and Clustering
Principal Component Analysis and ClusteringPrincipal Component Analysis and Clustering
Principal Component Analysis and Clustering
 
Discriminant analysis
Discriminant analysisDiscriminant analysis
Discriminant analysis
 
Regression analysis in R
Regression analysis in RRegression analysis in R
Regression analysis in R
 
Introduction to Machine Learning Classifiers
Introduction to Machine Learning ClassifiersIntroduction to Machine Learning Classifiers
Introduction to Machine Learning Classifiers
 
K mean-clustering algorithm
K mean-clustering algorithmK mean-clustering algorithm
K mean-clustering algorithm
 
Pca ppt
Pca pptPca ppt
Pca ppt
 
Fishers linear discriminant for dimensionality reduction.
Fishers linear discriminant for dimensionality reduction.Fishers linear discriminant for dimensionality reduction.
Fishers linear discriminant for dimensionality reduction.
 
Clustering
ClusteringClustering
Clustering
 
Analysis of variance ppt @ bec doms
Analysis of variance ppt @ bec domsAnalysis of variance ppt @ bec doms
Analysis of variance ppt @ bec doms
 
Presentation on K-Means Clustering
Presentation on K-Means ClusteringPresentation on K-Means Clustering
Presentation on K-Means Clustering
 
Correlation and regression
Correlation and regressionCorrelation and regression
Correlation and regression
 
Discriminant analysis group no. 4
Discriminant analysis  group no. 4Discriminant analysis  group no. 4
Discriminant analysis group no. 4
 
Introduction to Linear Discriminant Analysis
Introduction to Linear Discriminant AnalysisIntroduction to Linear Discriminant Analysis
Introduction to Linear Discriminant Analysis
 
Logistic regression
Logistic regressionLogistic regression
Logistic regression
 
Logistic regression
Logistic regressionLogistic regression
Logistic regression
 
Lect4 principal component analysis-I
Lect4 principal component analysis-ILect4 principal component analysis-I
Lect4 principal component analysis-I
 

Similar a Discriminant Analysis-lecture 8

Examining Differences in CES-D Measurement in The Health and Retirement Study
Examining Differences in CES-D Measurement in The Health and Retirement StudyExamining Differences in CES-D Measurement in The Health and Retirement Study
Examining Differences in CES-D Measurement in The Health and Retirement StudyAdam Perzynski, PhD
 
Basics of probability in statistical simulation and stochastic programming
Basics of probability in statistical simulation and stochastic programmingBasics of probability in statistical simulation and stochastic programming
Basics of probability in statistical simulation and stochastic programmingSSA KPI
 
Multi Level Modelling&Weights Workshop Kiel09
Multi Level Modelling&Weights Workshop Kiel09Multi Level Modelling&Weights Workshop Kiel09
Multi Level Modelling&Weights Workshop Kiel09egebhardt72
 
Linear Discriminant Analysis and Its Generalization
Linear Discriminant Analysis and Its GeneralizationLinear Discriminant Analysis and Its Generalization
Linear Discriminant Analysis and Its Generalization일상 온
 
U1.4-RVDistributions.ppt
U1.4-RVDistributions.pptU1.4-RVDistributions.ppt
U1.4-RVDistributions.pptSameeraasif2
 
T Test For Two Independent Samples
T Test For Two Independent SamplesT Test For Two Independent Samples
T Test For Two Independent Samplesshoffma5
 
SimpleLinearRegressionAnalysisWithExamples.ppt
SimpleLinearRegressionAnalysisWithExamples.pptSimpleLinearRegressionAnalysisWithExamples.ppt
SimpleLinearRegressionAnalysisWithExamples.pptAdnanAli861711
 
Linear regression.ppt
Linear regression.pptLinear regression.ppt
Linear regression.pptbranlymbunga1
 
Slideset Simple Linear Regression models.ppt
Slideset Simple Linear Regression models.pptSlideset Simple Linear Regression models.ppt
Slideset Simple Linear Regression models.pptrahulrkmgb09
 
lecture13.ppt
lecture13.pptlecture13.ppt
lecture13.pptarkian3
 
Classification of aedes adults mosquitoes in two distinct groups based on fis...
Classification of aedes adults mosquitoes in two distinct groups based on fis...Classification of aedes adults mosquitoes in two distinct groups based on fis...
Classification of aedes adults mosquitoes in two distinct groups based on fis...Alexander Decker
 
2012 mdsp pr02 1004
2012 mdsp pr02 10042012 mdsp pr02 1004
2012 mdsp pr02 1004nozomuhamada
 

Similar a Discriminant Analysis-lecture 8 (20)

Abu
AbuAbu
Abu
 
Examining Differences in CES-D Measurement in The Health and Retirement Study
Examining Differences in CES-D Measurement in The Health and Retirement StudyExamining Differences in CES-D Measurement in The Health and Retirement Study
Examining Differences in CES-D Measurement in The Health and Retirement Study
 
Bayes ML.ppt
Bayes ML.pptBayes ML.ppt
Bayes ML.ppt
 
Basics of probability in statistical simulation and stochastic programming
Basics of probability in statistical simulation and stochastic programmingBasics of probability in statistical simulation and stochastic programming
Basics of probability in statistical simulation and stochastic programming
 
Symmetrical2
Symmetrical2Symmetrical2
Symmetrical2
 
Multi Level Modelling&Weights Workshop Kiel09
Multi Level Modelling&Weights Workshop Kiel09Multi Level Modelling&Weights Workshop Kiel09
Multi Level Modelling&Weights Workshop Kiel09
 
Ez25936938
Ez25936938Ez25936938
Ez25936938
 
Linear Discriminant Analysis and Its Generalization
Linear Discriminant Analysis and Its GeneralizationLinear Discriminant Analysis and Its Generalization
Linear Discriminant Analysis and Its Generalization
 
U1.4-RVDistributions.ppt
U1.4-RVDistributions.pptU1.4-RVDistributions.ppt
U1.4-RVDistributions.ppt
 
T Test For Two Independent Samples
T Test For Two Independent SamplesT Test For Two Independent Samples
T Test For Two Independent Samples
 
SimpleLinearRegressionAnalysisWithExamples.ppt
SimpleLinearRegressionAnalysisWithExamples.pptSimpleLinearRegressionAnalysisWithExamples.ppt
SimpleLinearRegressionAnalysisWithExamples.ppt
 
Linear regression.ppt
Linear regression.pptLinear regression.ppt
Linear regression.ppt
 
lecture13.ppt
lecture13.pptlecture13.ppt
lecture13.ppt
 
lecture13.ppt
lecture13.pptlecture13.ppt
lecture13.ppt
 
Slideset Simple Linear Regression models.ppt
Slideset Simple Linear Regression models.pptSlideset Simple Linear Regression models.ppt
Slideset Simple Linear Regression models.ppt
 
lecture13.ppt
lecture13.pptlecture13.ppt
lecture13.ppt
 
lecture13.ppt
lecture13.pptlecture13.ppt
lecture13.ppt
 
Classification of aedes adults mosquitoes in two distinct groups based on fis...
Classification of aedes adults mosquitoes in two distinct groups based on fis...Classification of aedes adults mosquitoes in two distinct groups based on fis...
Classification of aedes adults mosquitoes in two distinct groups based on fis...
 
2012 mdsp pr02 1004
2012 mdsp pr02 10042012 mdsp pr02 1004
2012 mdsp pr02 1004
 
Elhabian lda09
Elhabian lda09Elhabian lda09
Elhabian lda09
 

Último

2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptxMaritesTamaniVerdade
 
SKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptx
SKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptxSKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptx
SKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptxAmanpreet Kaur
 
TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...
TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...
TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...Nguyen Thanh Tu Collection
 
Google Gemini An AI Revolution in Education.pptx
Google Gemini An AI Revolution in Education.pptxGoogle Gemini An AI Revolution in Education.pptx
Google Gemini An AI Revolution in Education.pptxDr. Sarita Anand
 
ICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptxICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptxAreebaZafar22
 
This PowerPoint helps students to consider the concept of infinity.
This PowerPoint helps students to consider the concept of infinity.This PowerPoint helps students to consider the concept of infinity.
This PowerPoint helps students to consider the concept of infinity.christianmathematics
 
Jamworks pilot and AI at Jisc (20/03/2024)
Jamworks pilot and AI at Jisc (20/03/2024)Jamworks pilot and AI at Jisc (20/03/2024)
Jamworks pilot and AI at Jisc (20/03/2024)Jisc
 
How to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptxHow to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptxCeline George
 
Wellbeing inclusion and digital dystopias.pptx
Wellbeing inclusion and digital dystopias.pptxWellbeing inclusion and digital dystopias.pptx
Wellbeing inclusion and digital dystopias.pptxJisc
 
HMCS Max Bernays Pre-Deployment Brief (May 2024).pptx
HMCS Max Bernays Pre-Deployment Brief (May 2024).pptxHMCS Max Bernays Pre-Deployment Brief (May 2024).pptx
HMCS Max Bernays Pre-Deployment Brief (May 2024).pptxEsquimalt MFRC
 
Salient Features of India constitution especially power and functions
Salient Features of India constitution especially power and functionsSalient Features of India constitution especially power and functions
Salient Features of India constitution especially power and functionsKarakKing
 
Introduction to Nonprofit Accounting: The Basics
Introduction to Nonprofit Accounting: The BasicsIntroduction to Nonprofit Accounting: The Basics
Introduction to Nonprofit Accounting: The BasicsTechSoup
 
Unit 3 Emotional Intelligence and Spiritual Intelligence.pdf
Unit 3 Emotional Intelligence and Spiritual Intelligence.pdfUnit 3 Emotional Intelligence and Spiritual Intelligence.pdf
Unit 3 Emotional Intelligence and Spiritual Intelligence.pdfDr Vijay Vishwakarma
 
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...ZurliaSoop
 
Key note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdfKey note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdfAdmir Softic
 
Holdier Curriculum Vitae (April 2024).pdf
Holdier Curriculum Vitae (April 2024).pdfHoldier Curriculum Vitae (April 2024).pdf
Holdier Curriculum Vitae (April 2024).pdfagholdier
 
REMIFENTANIL: An Ultra short acting opioid.pptx
REMIFENTANIL: An Ultra short acting opioid.pptxREMIFENTANIL: An Ultra short acting opioid.pptx
REMIFENTANIL: An Ultra short acting opioid.pptxDr. Ravikiran H M Gowda
 
Fostering Friendships - Enhancing Social Bonds in the Classroom
Fostering Friendships - Enhancing Social Bonds  in the ClassroomFostering Friendships - Enhancing Social Bonds  in the Classroom
Fostering Friendships - Enhancing Social Bonds in the ClassroomPooky Knightsmith
 
How to Manage Global Discount in Odoo 17 POS
How to Manage Global Discount in Odoo 17 POSHow to Manage Global Discount in Odoo 17 POS
How to Manage Global Discount in Odoo 17 POSCeline George
 

Último (20)

2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
2024-NATIONAL-LEARNING-CAMP-AND-OTHER.pptx
 
SKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptx
SKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptxSKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptx
SKILL OF INTRODUCING THE LESSON MICRO SKILLS.pptx
 
TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...
TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...
TỔNG ÔN TẬP THI VÀO LỚP 10 MÔN TIẾNG ANH NĂM HỌC 2023 - 2024 CÓ ĐÁP ÁN (NGỮ Â...
 
Google Gemini An AI Revolution in Education.pptx
Google Gemini An AI Revolution in Education.pptxGoogle Gemini An AI Revolution in Education.pptx
Google Gemini An AI Revolution in Education.pptx
 
ICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptxICT Role in 21st Century Education & its Challenges.pptx
ICT Role in 21st Century Education & its Challenges.pptx
 
This PowerPoint helps students to consider the concept of infinity.
This PowerPoint helps students to consider the concept of infinity.This PowerPoint helps students to consider the concept of infinity.
This PowerPoint helps students to consider the concept of infinity.
 
Jamworks pilot and AI at Jisc (20/03/2024)
Jamworks pilot and AI at Jisc (20/03/2024)Jamworks pilot and AI at Jisc (20/03/2024)
Jamworks pilot and AI at Jisc (20/03/2024)
 
How to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptxHow to setup Pycharm environment for Odoo 17.pptx
How to setup Pycharm environment for Odoo 17.pptx
 
Wellbeing inclusion and digital dystopias.pptx
Wellbeing inclusion and digital dystopias.pptxWellbeing inclusion and digital dystopias.pptx
Wellbeing inclusion and digital dystopias.pptx
 
HMCS Max Bernays Pre-Deployment Brief (May 2024).pptx
HMCS Max Bernays Pre-Deployment Brief (May 2024).pptxHMCS Max Bernays Pre-Deployment Brief (May 2024).pptx
HMCS Max Bernays Pre-Deployment Brief (May 2024).pptx
 
Mehran University Newsletter Vol-X, Issue-I, 2024
Mehran University Newsletter Vol-X, Issue-I, 2024Mehran University Newsletter Vol-X, Issue-I, 2024
Mehran University Newsletter Vol-X, Issue-I, 2024
 
Salient Features of India constitution especially power and functions
Salient Features of India constitution especially power and functionsSalient Features of India constitution especially power and functions
Salient Features of India constitution especially power and functions
 
Introduction to Nonprofit Accounting: The Basics
Introduction to Nonprofit Accounting: The BasicsIntroduction to Nonprofit Accounting: The Basics
Introduction to Nonprofit Accounting: The Basics
 
Unit 3 Emotional Intelligence and Spiritual Intelligence.pdf
Unit 3 Emotional Intelligence and Spiritual Intelligence.pdfUnit 3 Emotional Intelligence and Spiritual Intelligence.pdf
Unit 3 Emotional Intelligence and Spiritual Intelligence.pdf
 
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
Jual Obat Aborsi Hongkong ( Asli No.1 ) 085657271886 Obat Penggugur Kandungan...
 
Key note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdfKey note speaker Neum_Admir Softic_ENG.pdf
Key note speaker Neum_Admir Softic_ENG.pdf
 
Holdier Curriculum Vitae (April 2024).pdf
Holdier Curriculum Vitae (April 2024).pdfHoldier Curriculum Vitae (April 2024).pdf
Holdier Curriculum Vitae (April 2024).pdf
 
REMIFENTANIL: An Ultra short acting opioid.pptx
REMIFENTANIL: An Ultra short acting opioid.pptxREMIFENTANIL: An Ultra short acting opioid.pptx
REMIFENTANIL: An Ultra short acting opioid.pptx
 
Fostering Friendships - Enhancing Social Bonds in the Classroom
Fostering Friendships - Enhancing Social Bonds  in the ClassroomFostering Friendships - Enhancing Social Bonds  in the Classroom
Fostering Friendships - Enhancing Social Bonds in the Classroom
 
How to Manage Global Discount in Odoo 17 POS
How to Manage Global Discount in Odoo 17 POSHow to Manage Global Discount in Odoo 17 POS
How to Manage Global Discount in Odoo 17 POS
 

Discriminant Analysis-lecture 8

  • 1. 4/30/2012 Linear Discriminant Analysis  Proposed by Fisher (1936) for classifying an observation into one of two possible groups based on many measurements x1,x2,…xp.  Seek a linear transformation of the variables Y=w1x1+w2x2+..+wpxp + a constant 1
  • 2. 4/30/2012 Linear Discriminant Analysis  Discriminant analysis – creates an equation which will minimize the possibility of misclassifying cases into their respective groups or categories. The purposes of discriminant analysis (DA)  Discriminant Function Analysis (DA) undertakes the same task as multiple linear regression by predicting an outcome.  However, multiple linear regression is limited to cases where the dependent is numerical  But many interesting variables are categorical, 2
  • 3. 4/30/2012  The objective of DA is to perform dimensionality reduction while preserving as much of the class discriminatory information as possible  Assume we have a set of D-dimensional samples {x 1, x2, …, xN}, N1 of which belong to class ω1, and N2 to class ω2.  We seek to obtain a scalar y by projecting the samples x onto a line y = wTx •The top two distributions overlap too much and do not discriminate too well compared to the bottom set. •Misclassification will be minimal in the lower pair, •whereas many will be misclassified in the top pair. 3
  • 4. 4/30/2012 Linear Discriminant Analysis Assume variance matrices equal Classify the item x at hand to one of J groups based on measurements on p predictors. Rule: Assign x to group j that has the closest mean j = 1, 2, …, J Distance Measure: Mahalanobis Distance. Linear Discriminant Analysis Distance Measure: For j = 1, 2, …, J, compute d j  x    x  x j Spl  x  x j T 1 Assign x to the group for which dj is minimum S is the pooled estimate of the covariance pl matrix 4
  • 5. 4/30/2012 …or equivalently, assign x to the group for which L x   x S 1 1 1 x T T j j pl 2 x S x j pl j is a maximum. (Notice the linear form of the equation!) Linear Discriminant Analysis …optimal if…. • Multivariate normal distribution for the observation in each of the groups • Equal covariance matrix for all groups • Equal prior probability for each group • Equal costs for misclassification 5
  • 6. 4/30/2012 Relaxing the assumption of equal prior probabilities… L x   ln p j  x S T 1 1 1 x T j j pl 2 x S x j pl j pj being the prior probability for the jth group. Relaxing the assumption of equal covariance matrices… 1 Q  x   ln p j  ln S j j 2  x  x  S x  x j  T 1 j j result?…Quadratic Discriminant Analysis 6
  • 7. 4/30/2012 Quadratic Discriminant Analysis Rule: assign to group j if Q x  j is the largest. Optimal if the J groups of measurements are multivariate normal Other Extensions & Related Methods Relaxing the assumption of normality… Kernel density based LDA and QDA Other extensions….. Regularized discriminant analysis Penalized discriminant analysis Flexible discriminant analysis 7
  • 8. 4/30/2012 Evaluations of the Methods Classification Table (confusion matrix) Predicted group Actual group Number of observations A B A nA n11 n12 B nB n21 n22 Evaluations of the Methods Apparent Error Rate (APER): # misclassified APER = Total # of cases ….underestimates the actual error rate. Improved estimate of APER: Holdout Method or cross validation 8
  • 9. 4/30/2012 Fisher's iris dataset •The data were collected by Anderson and used by Fisher to formulate the linear discriminant analysis (LDA or DA). •The dataset gives the measurements in centimeters of the following variables: 1- sepal length, 2- sepal width, 3- petal length, and 4- petal width, this for 50 fowers from each of the 3 species of iris considered. •The species considered are Iris setosa, versicolor, and virginica setosa versicolor virginica 9
  • 10. 4/30/2012 An Example: Fisher’s Iris Data Actual Number of Predicted Group Observations Group Setosa Versicolo Virginica r Setosa 50 50 0 0 Versicolor 50 0 48 2 Virginica 50 0 1 49 Table 1: Linear Discriminant Analysis (APER = 0.0200) An Example: Fisher’s Iris Data Actual Number of Predicted Group Observations Group Setosa Versicolo Virginica r Setosa 50 50 0 0 Versicolor 50 0 47 3 Virginica 50 0 1 49 Table 1: Quadratic Discriminant Analysis (APER = 0.0267) 10
  • 11. 4/30/2012 An Example: Fisher’s Iris Data 2.5 v v v v v v v v v v v v v v v v v 2.0 v v v v v v v v v v v v v v c v v c v c v c c 1.5 c v c c v c c c c v c c c c c c Petal Width c c c c c c c c c c c c 1.0 c c c c c c s 0.5 s s s s s s s s s s s s s s s s s s s s s s s s s s s 2. 0 2. 5 3. 0 3. 5 4. 0 S Wi t epal dh An Example: Fisher’s Iris Data 2.5 o v o v o v o v o v o v o v o v o v o v o v o v o v o v o v o v o v 2.0 o v o v o v o v o v o v o v o v o v o v o v o v o v o v o c v o v x c o c o v x c x c 1.5 x o c v x c x c v x c x c x c x c o v x c x c x c x c x c x c Petal Width x c x c x c x c x c x c x c x c x c x c x c x c 1.0 x c x c x c x c x c x c + s 0.5 + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s + s 2. 0 2. 5 3. 0 3. 5 4. 0 S Wi t epal dh 11
  • 12. 4/30/2012 Summary LDA is a powerful tool available for classification. Widely implemented through various software Theoretical properties well researched 12