Enviar pesquisa
Carregar
Information Gain
•
1 gostou
•
2,562 visualizações
G
guest32311f
Seguir
Tecnologia
Educação
Denunciar
Compartilhar
Denunciar
Compartilhar
1 de 10
Baixar agora
Baixar para ler offline
Recomendados
2013-1 Machine Learning Lecture 02 - Andrew Moore: Entropy
2013-1 Machine Learning Lecture 02 - Andrew Moore: Entropy
Dongseo University
Csr2011 june15 09_30_shen
Csr2011 june15 09_30_shen
CSR2011
Lesson19 Maximum And Minimum Values 034 Slides
Lesson19 Maximum And Minimum Values 034 Slides
Matthew Leingang
Multimodal pattern matching algorithms and applications
Multimodal pattern matching algorithms and applications
Xavier Anguera
Fourier Transforms
Fourier Transforms
Arvind Devaraj
Fec512.02
Fec512.02
Orhan Erdem
Time Machine session @ ICME 2012 - DTW's New Youth
Time Machine session @ ICME 2012 - DTW's New Youth
Xavier Anguera
Probability Density Functions
Probability Density Functions
guestb86588
Recomendados
2013-1 Machine Learning Lecture 02 - Andrew Moore: Entropy
2013-1 Machine Learning Lecture 02 - Andrew Moore: Entropy
Dongseo University
Csr2011 june15 09_30_shen
Csr2011 june15 09_30_shen
CSR2011
Lesson19 Maximum And Minimum Values 034 Slides
Lesson19 Maximum And Minimum Values 034 Slides
Matthew Leingang
Multimodal pattern matching algorithms and applications
Multimodal pattern matching algorithms and applications
Xavier Anguera
Fourier Transforms
Fourier Transforms
Arvind Devaraj
Fec512.02
Fec512.02
Orhan Erdem
Time Machine session @ ICME 2012 - DTW's New Youth
Time Machine session @ ICME 2012 - DTW's New Youth
Xavier Anguera
Probability Density Functions
Probability Density Functions
guestb86588
Team meeting 100325
Team meeting 100325
Yi-Hsin Liu
Lesson 18: Maximum and Minimum Vaues
Lesson 18: Maximum and Minimum Vaues
Matthew Leingang
Lesson 18: Maximum and Minimum Vaues
Lesson 18: Maximum and Minimum Vaues
Matthew Leingang
Discrete Models in Computer Vision
Discrete Models in Computer Vision
Yap Wooi Hen
Minimax Rates for Homology Inference
Minimax Rates for Homology Inference
Don Sheehy
Midterm I Review
Midterm I Review
Matthew Leingang
Backpropagation in Convolutional Neural Network
Backpropagation in Convolutional Neural Network
Hiroshi Kuwajima
Bayesian Subset Simulation
Bayesian Subset Simulation
Julien Bect
Sparse autoencoder
Sparse autoencoder
Devashish Patel
Accelerometers to Augmented Reality
Accelerometers to Augmented Reality
jblocksom
Introducing Copula to Risk Management Presentation
Introducing Copula to Risk Management Presentation
Eva Li
Applying the derivative
Applying the derivative
Inarotul Faiza
Lesson 16: Derivatives of Exponential and Logarithmic Functions
Lesson 16: Derivatives of Exponential and Logarithmic Functions
Matthew Leingang
Digital Communication: Information Theory
Digital Communication: Information Theory
Dr. Sanjay M. Gulhane
Introduction to Tensorflow
Introduction to Tensorflow
Tzar Umang
Lesson 22: Optimization I (Section 4 version)
Lesson 22: Optimization I (Section 4 version)
Matthew Leingang
New Bounds on the Size of Optimal Meshes
New Bounds on the Size of Optimal Meshes
Don Sheehy
Nodebox for Data Visualization
Nodebox for Data Visualization
Lynn Cherny
Lesson 8: Derivatives of Polynomials and Exponential functions
Lesson 8: Derivatives of Polynomials and Exponential functions
Matthew Leingang
Probability density in data mining and coveriance
Probability density in data mining and coveriance
udhayax793
Lecture7 channel capacity
Lecture7 channel capacity
Frank Katta
advance coding techniques - probability
advance coding techniques - probability
YaseenMo
Mais conteúdo relacionado
Mais procurados
Team meeting 100325
Team meeting 100325
Yi-Hsin Liu
Lesson 18: Maximum and Minimum Vaues
Lesson 18: Maximum and Minimum Vaues
Matthew Leingang
Lesson 18: Maximum and Minimum Vaues
Lesson 18: Maximum and Minimum Vaues
Matthew Leingang
Discrete Models in Computer Vision
Discrete Models in Computer Vision
Yap Wooi Hen
Minimax Rates for Homology Inference
Minimax Rates for Homology Inference
Don Sheehy
Midterm I Review
Midterm I Review
Matthew Leingang
Backpropagation in Convolutional Neural Network
Backpropagation in Convolutional Neural Network
Hiroshi Kuwajima
Bayesian Subset Simulation
Bayesian Subset Simulation
Julien Bect
Sparse autoencoder
Sparse autoencoder
Devashish Patel
Accelerometers to Augmented Reality
Accelerometers to Augmented Reality
jblocksom
Introducing Copula to Risk Management Presentation
Introducing Copula to Risk Management Presentation
Eva Li
Applying the derivative
Applying the derivative
Inarotul Faiza
Lesson 16: Derivatives of Exponential and Logarithmic Functions
Lesson 16: Derivatives of Exponential and Logarithmic Functions
Matthew Leingang
Digital Communication: Information Theory
Digital Communication: Information Theory
Dr. Sanjay M. Gulhane
Introduction to Tensorflow
Introduction to Tensorflow
Tzar Umang
Lesson 22: Optimization I (Section 4 version)
Lesson 22: Optimization I (Section 4 version)
Matthew Leingang
New Bounds on the Size of Optimal Meshes
New Bounds on the Size of Optimal Meshes
Don Sheehy
Nodebox for Data Visualization
Nodebox for Data Visualization
Lynn Cherny
Lesson 8: Derivatives of Polynomials and Exponential functions
Lesson 8: Derivatives of Polynomials and Exponential functions
Matthew Leingang
Mais procurados
(19)
Team meeting 100325
Team meeting 100325
Lesson 18: Maximum and Minimum Vaues
Lesson 18: Maximum and Minimum Vaues
Lesson 18: Maximum and Minimum Vaues
Lesson 18: Maximum and Minimum Vaues
Discrete Models in Computer Vision
Discrete Models in Computer Vision
Minimax Rates for Homology Inference
Minimax Rates for Homology Inference
Midterm I Review
Midterm I Review
Backpropagation in Convolutional Neural Network
Backpropagation in Convolutional Neural Network
Bayesian Subset Simulation
Bayesian Subset Simulation
Sparse autoencoder
Sparse autoencoder
Accelerometers to Augmented Reality
Accelerometers to Augmented Reality
Introducing Copula to Risk Management Presentation
Introducing Copula to Risk Management Presentation
Applying the derivative
Applying the derivative
Lesson 16: Derivatives of Exponential and Logarithmic Functions
Lesson 16: Derivatives of Exponential and Logarithmic Functions
Digital Communication: Information Theory
Digital Communication: Information Theory
Introduction to Tensorflow
Introduction to Tensorflow
Lesson 22: Optimization I (Section 4 version)
Lesson 22: Optimization I (Section 4 version)
New Bounds on the Size of Optimal Meshes
New Bounds on the Size of Optimal Meshes
Nodebox for Data Visualization
Nodebox for Data Visualization
Lesson 8: Derivatives of Polynomials and Exponential functions
Lesson 8: Derivatives of Polynomials and Exponential functions
Semelhante a Information Gain
Probability density in data mining and coveriance
Probability density in data mining and coveriance
udhayax793
Lecture7 channel capacity
Lecture7 channel capacity
Frank Katta
advance coding techniques - probability
advance coding techniques - probability
YaseenMo
Losseless
Losseless
anithabalaprabhu
Probabilistic systems exam help
Probabilistic systems exam help
Statistics Exam Help
Probabilistic systems assignment help
Probabilistic systems assignment help
Statistics Assignment Help
Tele4653 l9
Tele4653 l9
Vin Voro
Lecture 2
Lecture 2
Shravan Vasishth
Probability distribution for Dummies
Probability distribution for Dummies
Balaji P
Morphing Image
Morphing Image
Copen Hagen
Dcs unit 2
Dcs unit 2
Anil Nigam
Unit 5.pdf
Unit 5.pdf
BALASHANMUGAVADIVUPM
Maximum Likelihood Estimation
Maximum Likelihood Estimation
guestfee8698
PAC Learning
PAC Learning
guestfee8698
My7class
My7class
ketan533
Dirichlet processes and Applications
Dirichlet processes and Applications
Saurav Jha
Moment-Generating Functions and Reproductive Properties of Distributions
Moment-Generating Functions and Reproductive Properties of Distributions
IJSRED
IJSRED-V2I5P56
IJSRED-V2I5P56
IJSRED
Predicting Real-valued Outputs: An introduction to regression
Predicting Real-valued Outputs: An introduction to regression
guestfee8698
7주차
7주차
Kookmin University
Semelhante a Information Gain
(20)
Probability density in data mining and coveriance
Probability density in data mining and coveriance
Lecture7 channel capacity
Lecture7 channel capacity
advance coding techniques - probability
advance coding techniques - probability
Losseless
Losseless
Probabilistic systems exam help
Probabilistic systems exam help
Probabilistic systems assignment help
Probabilistic systems assignment help
Tele4653 l9
Tele4653 l9
Lecture 2
Lecture 2
Probability distribution for Dummies
Probability distribution for Dummies
Morphing Image
Morphing Image
Dcs unit 2
Dcs unit 2
Unit 5.pdf
Unit 5.pdf
Maximum Likelihood Estimation
Maximum Likelihood Estimation
PAC Learning
PAC Learning
My7class
My7class
Dirichlet processes and Applications
Dirichlet processes and Applications
Moment-Generating Functions and Reproductive Properties of Distributions
Moment-Generating Functions and Reproductive Properties of Distributions
IJSRED-V2I5P56
IJSRED-V2I5P56
Predicting Real-valued Outputs: An introduction to regression
Predicting Real-valued Outputs: An introduction to regression
7주차
7주차
Último
[BuildWithAI] Introduction to Gemini.pdf
[BuildWithAI] Introduction to Gemini.pdf
Sandro Moreira
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
Martijn de Jong
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...
Angeliki Cooney
Apidays New York 2024 - APIs in 2030: The Risk of Technological Sleepwalk by ...
Apidays New York 2024 - APIs in 2030: The Risk of Technological Sleepwalk by ...
apidays
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
apidays
Exploring Multimodal Embeddings with Milvus
Exploring Multimodal Embeddings with Milvus
Zilliz
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
Remote DBA Services
CNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In Pakistan
danishmna97
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
ThousandEyes
Ransomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdf
Overkill Security
AXA XL - Insurer Innovation Award Americas 2024
AXA XL - Insurer Innovation Award Americas 2024
The Digital Insurer
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
DianaGray10
presentation ICT roal in 21st century education
presentation ICT roal in 21st century education
jfdjdjcjdnsjd
DEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
DEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
UiPathCommunity
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
The Digital Insurer
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
Product Anonymous
DBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor Presentation
Dropbox
Rising Above_ Dubai Floods and the Fortitude of Dubai International Airport.pdf
Rising Above_ Dubai Floods and the Fortitude of Dubai International Airport.pdf
Orbitshub
Spring Boot vs Quarkus the ultimate battle - DevoxxUK
Spring Boot vs Quarkus the ultimate battle - DevoxxUK
Jago de Vreede
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Zilliz
Último
(20)
[BuildWithAI] Introduction to Gemini.pdf
[BuildWithAI] Introduction to Gemini.pdf
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...
Biography Of Angeliki Cooney | Senior Vice President Life Sciences | Albany, ...
Apidays New York 2024 - APIs in 2030: The Risk of Technological Sleepwalk by ...
Apidays New York 2024 - APIs in 2030: The Risk of Technological Sleepwalk by ...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Apidays New York 2024 - Accelerating FinTech Innovation by Vasa Krishnan, Fin...
Exploring Multimodal Embeddings with Milvus
Exploring Multimodal Embeddings with Milvus
Strategies for Landing an Oracle DBA Job as a Fresher
Strategies for Landing an Oracle DBA Job as a Fresher
CNIC Information System with Pakdata Cf In Pakistan
CNIC Information System with Pakdata Cf In Pakistan
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
Ransomware_Q4_2023. The report. [EN].pdf
Ransomware_Q4_2023. The report. [EN].pdf
AXA XL - Insurer Innovation Award Americas 2024
AXA XL - Insurer Innovation Award Americas 2024
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
Connector Corner: Accelerate revenue generation using UiPath API-centric busi...
presentation ICT roal in 21st century education
presentation ICT roal in 21st century education
DEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
DEV meet-up UiPath Document Understanding May 7 2024 Amsterdam
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
DBX First Quarter 2024 Investor Presentation
DBX First Quarter 2024 Investor Presentation
Rising Above_ Dubai Floods and the Fortitude of Dubai International Airport.pdf
Rising Above_ Dubai Floods and the Fortitude of Dubai International Airport.pdf
Spring Boot vs Quarkus the ultimate battle - DevoxxUK
Spring Boot vs Quarkus the ultimate battle - DevoxxUK
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Emergent Methods: Multi-lingual narrative tracking in the news - real-time ex...
Information Gain
1.
Note to other
teachers and users of these slides. Andrew would be delighted if you found this source material useful in giving your own lectures. Feel free to use these slides verbatim, or to modify them to fit your own needs. PowerPoint originals are available. If you make use of a significant portion of these slides in your own lecture, please include this message, or the following link to the source repository of Andrew’s tutorials: http://www.cs.cmu.edu/~awm/tutorials . Comments and corrections gratefully received. Information Gain Andrew W. Moore Professor School of Computer Science Carnegie Mellon University www.cs.cmu.edu/~awm awm@cs.cmu.edu 412-268-7599 Copyright © 2001, 2003, Andrew W. Moore Bits You are watching a set of independent random samples of X You see that X has four possible values P(X=A) = 1/4 P(X=B) = 1/4 P(X=C) = 1/4 P(X=D) = 1/4 So you might see: BAACBADCDADDDA… You transmit data over a binary serial link. You can encode each reading with two bits (e.g. A = 00, B = 01, C = 10, D = 11) 0100001001001110110011111100… Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 2 1
2.
Fewer Bits Someone
tells you that the probabilities are not equal P(X=A) = 1/2 P(X=B) = 1/4 P(X=C) = 1/8 P(X=D) = 1/8 It’s possible… …to invent a coding for your transmission that only uses 1.75 bits on average per symbol. How? Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 3 Fewer Bits Someone tells you that the probabilities are not equal P(X=A) = 1/2 P(X=B) = 1/4 P(X=C) = 1/8 P(X=D) = 1/8 It’s possible… …to invent a coding for your transmission that only uses 1.75 bits on average per symbol. How? A 0 B 10 C 110 D 111 (This is just one of several ways) Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 4 2
3.
Fewer Bits Suppose there
are three equally likely values… P(X=A) = 1/3 P(X=B) = 1/3 P(X=C) = 1/3 Here’s a naïve coding, costing 2 bits per symbol A 00 B 01 C 10 Can you think of a coding that would need only 1.6 bits per symbol on average? In theory, it can in fact be done with 1.58496 bits per symbol. Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 5 General Case Suppose X can have one of m values… V1, V2, … Vm P(X=V1) = p1 P(X=V2) = p2 …. P(X=Vm) = pm What’s the smallest possible number of bits, on average, per symbol, needed to transmit a stream of symbols drawn from X’s distribution? It’s H ( X ) = − p1 log 2 p1 − p2 log 2 p2 − K − pm log 2 pm m = −∑ p j log 2 p j j =1 H(X) = The entropy of X • “High Entropy” means X is from a uniform (boring) distribution • “Low Entropy” means X is from varied (peaks and valleys) distribution Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 6 3
4.
General Case Suppose X
can have one of m values… V1, V2, … Vm P(X=V1) = p1 P(X=V2) = p2 …. P(X=Vm) = pm A histogram of the What’s the smallest possible number of frequency average, per bits, on distribution of symbol, needed to transmit a stream values of X would have of symbols drawn from A histogram of the X’s distribution? It’s many lows and one or frequency distribution of two highs H ( X ) = − p1 log 2would be flat 2 p2 − K − pm log 2 pm values of X p − p log 1 2 m = −∑ p j log 2 p j j =1 H(X) = The entropy of X • “High Entropy” means X is from a uniform (boring) distribution • “Low Entropy” means X is from varied (peaks and valleys) distribution Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 7 General Case Suppose X can have one of m values… V1, V2, … Vm P(X=V1) = p1 P(X=V2) = p2 …. P(X=Vm) = pm A histogram of the What’s the smallest possible number of frequency average, per bits, on distribution of symbol, needed to transmit a stream values of X would have of symbols drawn from A histogram of the X’s distribution? It’s many lows and one or frequency distribution of two highs H ( X ) = − p1 log 2would be flat 2 p2 − K − pm log 2 pm values of X p − p log 1 2 m = −∑ p ..and sop j values j log 2 the ..and so the values sampled from it would j =1 sampled from it would be all over the place be more predictable H(X) = The entropy of X • “High Entropy” means X is from a uniform (boring) distribution • “Low Entropy” means X is from varied (peaks and valleys) distribution Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 8 4
5.
Entropy in a
nut-shell Low Entropy High Entropy Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 9 Entropy in a nut-shell Low Entropy High Entropy ..the values (locations of ..the values (locations soup) unpredictable... of soup) sampled almost uniformly sampled entirely from within throughout our dining room the soup bowl Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 10 5
6.
Specific Conditional Entropy
H(Y|X=v) Suppose I’m trying to predict output Y and I have input X X = College Major Let’s assume this reflects the true probabilities Y = Likes “Gladiator” E.G. From this data we estimate X Y • P(LikeG = Yes) = 0.5 Math Yes History No • P(Major = Math & LikeG = No) = 0.25 CS Yes • P(Major = Math) = 0.5 Math No • P(LikeG = Yes | Major = History) = 0 Math No Note: CS Yes • H(X) = 1.5 History No •H(Y) = 1 Math Yes Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 11 Specific Conditional Entropy H(Y|X=v) Definition of Specific Conditional X = College Major Entropy: Y = Likes “Gladiator” H(Y |X=v) = The entropy of Y among only those records in which X Y X has value v Math Yes History No CS Yes Math No Math No CS Yes History No Math Yes Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 12 6
7.
Specific Conditional Entropy
H(Y|X=v) Definition of Specific Conditional X = College Major Entropy: Y = Likes “Gladiator” H(Y |X=v) = The entropy of Y among only those records in which X Y X has value v Math Yes Example: History No CS Yes • H(Y|X=Math) = 1 Math No • H(Y|X=History) = 0 Math No • H(Y|X=CS) = 0 CS Yes History No Math Yes Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 13 Conditional Entropy H(Y|X) X = College Major Definition of Conditional Entropy: Y = Likes “Gladiator” H(Y |X) = The average specific X Y conditional entropy of Y Math Yes = if you choose a record at random what History No will be the conditional entropy of Y, CS Yes conditioned on that row’s value of X Math No Math No = Expected number of bits to transmit Y if CS Yes both sides will know the value of X History No = Σj Prob(X=vj) H(Y | X = vj) Math Yes Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 14 7
8.
Conditional Entropy
Definition of Conditional Entropy: X = College Major Y = Likes “Gladiator” H(Y|X) = The average conditional entropy of Y = ΣjProb(X=vj) H(Y | X = vj) X Y Example: Math Yes History No vj Prob(X=vj) H(Y | X = vj) CS Yes Math 0.5 1 Math No History 0.25 0 Math No CS 0.25 0 CS Yes History No H(Y|X) = 0.5 * 1 + 0.25 * 0 + 0.25 * 0 = 0.5 Math Yes Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 15 Information Gain Definition of Information Gain: X = College Major Y = Likes “Gladiator” IG(Y|X) = I must transmit Y. How many bits on average would it save me if both ends of the line knew X? X Y IG(Y|X) = H(Y) - H(Y | X) Math Yes History No Example: CS Yes Math No • H(Y) = 1 Math No • H(Y|X) = 0.5 CS Yes • Thus IG(Y|X) = 1 – 0.5 = 0.5 History No Math Yes Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 16 8
9.
Information Gain Example Copyright
© 2001, 2003, Andrew W. Moore Information Gain: Slide 17 Another example Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 18 9
10.
Relative Information Gain
Definition of Relative Information X = College Major Y = Likes “Gladiator” Gain: RIG(Y|X) = I must transmit Y, what fraction of the bits on average would it save me if both ends of the line X Y knew X? Math Yes RIG(Y|X) = H(Y) - H(Y | X) / H(Y) History No CS Yes Example: Math No H(Y|X) = 0.5 Math No • CS Yes H(Y) = 1 • History No Thus IG(Y|X) = (1 – 0.5)/1 = 0.5 • Math Yes Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 19 What is Information Gain used for? Suppose you are trying to predict whether someone is going live past 80 years. From historical data you might find… •IG(LongLife | HairColor) = 0.01 •IG(LongLife | Smoker) = 0.2 •IG(LongLife | Gender) = 0.25 •IG(LongLife | LastDigitOfSSN) = 0.00001 IG tells you how interesting a 2-d contingency table is going to be. Copyright © 2001, 2003, Andrew W. Moore Information Gain: Slide 20 10
Baixar agora