Feature Selection in Machine
Learning
Presented by
R.Sridevi
Assistant Professor,
Department of Computer Science,
Tagore Govt.Arts & Science College,
Puducherry.
Agenda
Feature Selection
Study on Feature Selection Methods
Feature Selection in Machine Learning
Application of Feature Selection
Conclusion
Feature Selection
 Technique for identifying the most important features for learning.
 Increases learning performance by reducing the dimensionality of
data.
Subset
Validation
Filter Approach
 Filtering methods are independent of the induction algorithm.
 Evaluate each feature individually based on its correlation with the target function.
 Filter out irrelevant attributes before induction occurs
Input Features Feature Subset Selection Induction Algorithm
Wrapper Approach
Input Features Induction Algorithm
Wrapper Approach
Feature Subset Evaluation
Feature Subset Search
Induction Algorithm
A generic approach for feature selection occurs outside the basic
induction method but uses that method as a subroutine.
Searches the same space of feature subsets as filter methods, but it
evaluates alternative sets by running some induction algorithm on the
training data.
Relevant, Irrelevant & Redundant Features
 purpose of a Feature Selection is to identify relevant features according to
a definition of relevance.
 Relevance with respect to Target
 Strong Relevance to sample
 Weak Relevance to the sample
 presence of irrelevant attributes should considerably slow the rate of
learning.
 Redundant features exists whenever a feature can take the role of another.
Existing Feature Selection Methods
Feature weighting/ ranking algorithms.
Floating Search Methods in Feature Selection.
Mutual Information Based Algorithms.
Correlation based Algorithms.
Feature Interaction based Algorithms.
Categorization of Feature Selection
Methods
FSA Supporting Aspects
Relief, Relief-F,ABB, LVF Noise Tolerant
FOCUS, FCBF, Consistency,
mRMR
Eliminate Redundant Features
Filter Model based on GA Eliminate Irrelevant Features
MIFS, FSBAR, ABB,
CBFS,FRFS,IWFS
Eliminate Redundant & Irrelevant
Features
SAGA, INTERACT,
FSBAR,FRFS,IWFS Feature Interaction
Neural Network based Machine Learning
Supervised-Learning NN
Feed-Forward NN
 Perceptron
 Backpropagation (BP)
 LearningVector Quantization (LVQ)
Recurrent NN
 Fuzzy Cognitive Map (FCM)
 Boltzmann Machine (BM)
Unsupervised-Learning NN
Feed-Forward NN
 Learning Matrix(LM)
 Fuzzy Associative Memory (FAM)
Recurrent NN
 Binary Adaptive ResonanceTheory (ART1)
 Kohonen Self- Organizing Map (SOM)
Applications of Feature Selection
Text Mining
Image Processing
Bio- Informatics
Industrial Applications
Conclusion
Feature Selection improves accuracy during classification.
Enables the machine learning algorithm to train faster.
Avoid the curse of high dimensionality.
ThankYou

Bu-Refresher course PRESENTATION NEW.pptx

  • 1.
    Feature Selection inMachine Learning Presented by R.Sridevi Assistant Professor, Department of Computer Science, Tagore Govt.Arts & Science College, Puducherry.
  • 2.
    Agenda Feature Selection Study onFeature Selection Methods Feature Selection in Machine Learning Application of Feature Selection Conclusion
  • 3.
    Feature Selection  Techniquefor identifying the most important features for learning.  Increases learning performance by reducing the dimensionality of data. Subset Validation
  • 4.
    Filter Approach  Filteringmethods are independent of the induction algorithm.  Evaluate each feature individually based on its correlation with the target function.  Filter out irrelevant attributes before induction occurs Input Features Feature Subset Selection Induction Algorithm
  • 5.
    Wrapper Approach Input FeaturesInduction Algorithm Wrapper Approach Feature Subset Evaluation Feature Subset Search Induction Algorithm A generic approach for feature selection occurs outside the basic induction method but uses that method as a subroutine. Searches the same space of feature subsets as filter methods, but it evaluates alternative sets by running some induction algorithm on the training data.
  • 6.
    Relevant, Irrelevant &Redundant Features  purpose of a Feature Selection is to identify relevant features according to a definition of relevance.  Relevance with respect to Target  Strong Relevance to sample  Weak Relevance to the sample  presence of irrelevant attributes should considerably slow the rate of learning.  Redundant features exists whenever a feature can take the role of another.
  • 7.
    Existing Feature SelectionMethods Feature weighting/ ranking algorithms. Floating Search Methods in Feature Selection. Mutual Information Based Algorithms. Correlation based Algorithms. Feature Interaction based Algorithms.
  • 8.
    Categorization of FeatureSelection Methods FSA Supporting Aspects Relief, Relief-F,ABB, LVF Noise Tolerant FOCUS, FCBF, Consistency, mRMR Eliminate Redundant Features Filter Model based on GA Eliminate Irrelevant Features MIFS, FSBAR, ABB, CBFS,FRFS,IWFS Eliminate Redundant & Irrelevant Features SAGA, INTERACT, FSBAR,FRFS,IWFS Feature Interaction
  • 9.
    Neural Network basedMachine Learning Supervised-Learning NN Feed-Forward NN  Perceptron  Backpropagation (BP)  LearningVector Quantization (LVQ) Recurrent NN  Fuzzy Cognitive Map (FCM)  Boltzmann Machine (BM) Unsupervised-Learning NN Feed-Forward NN  Learning Matrix(LM)  Fuzzy Associative Memory (FAM) Recurrent NN  Binary Adaptive ResonanceTheory (ART1)  Kohonen Self- Organizing Map (SOM)
  • 10.
    Applications of FeatureSelection Text Mining Image Processing Bio- Informatics Industrial Applications
  • 11.
    Conclusion Feature Selection improvesaccuracy during classification. Enables the machine learning algorithm to train faster. Avoid the curse of high dimensionality.
  • 12.