Linear classifier MAXIMUM ENTROPY. Linear regression. Logistic regression 11/3/11. f 1
|
|
- Johnathan Campbell
- 8 years ago
- Views:
Transcription
1 Liear classifier A liear classifier predicts the label based o a weighted, liear combiatio of the features predictio = w 0 + w 1 f 1 + w 2 f w m f m For two classes, a liear classifier ca be viewed as a plae (hyperplae i the feature space MAXIMUM ENTROPY David Kauchak CS457, Sprig 2011 Some material derived from Jaso Eiser f 2 f 3 f 1 Liear regressio Logistic regressio Predict the respose based o a weighted, liear combiatio of the features log P(1 1 P(1 = w 0 + w 1 + w w m x m h( f = w 0 + w 1 f 1 + w 2 f w m f m real value weights Lear weights by miimizig the square error o the traiig data # error(h = (y i (w 0 + w 1 f 1 + w 2 f w m f m 2 P(1 1 P(1 = ew 0 +w 1 +w w mx m P(1 = (1 P(1 e w 0+w 1+w w mx m 1 P(1 = 1+ e (w 0 +w 1 +w w m x m 1
2 Logistic fuctio 1 logistic = 1+ e x Logistic regressio Fid the best fit of the data based o a logistic Logistic regressio 3 views of logistic regressio How would we classify examples oce we had a traied model? log P(1 1 P(1 = w 0 + w 1 + w w m x m liear classifier log P(1 1 P(1 = w 0 + w 1 + w w m x m If the sum > 0 the p(1/p(0 > 1, so positive if the sum < 0 the p(1/p(0 < 1, so egative Still a liear classifier (decisio boudary is a lie ew 0 +w 1 +w w m x m P(1 = 1+ e w 0 +w 1 +w w mx m 1 P(1 = 1+ e (w 0 +w 1 +w w m x m expoetial model (log-liear model logistic 2
3 Logistic regressio Fid the best fit of the data based o a logistic fuctio Traiig logistic regressio models How should we lear the parameters for logistic regressio (i.e. the w s? log P(1 1 P(1 = w 0 + w 1 + w w m x m parameters 1 P(1 = 1+ e (w 0 +w 1 +w w mx m Traiig logistic regressio models Idea 1: miimize the squared error (like liear regressio Ay problems? A digressio Why is this called Maximum Likelihood Estimatio (MLE? Parsed seteces Grammar log P(1 1 P(1 = w 0 + w 1 + w w m x m # error(h = (y i h( f i 2 We do t kow what the actual probability values are Learig/ Traiig cout( #$ P( #$ = cout( S NP VP 0.9 S VP 0.1 NP Det A N 0.5 NP NP PP 0.3 NP PropN 0.2 A ε 0.6 A Adj A 0.4 PP Prep NP 1.0 VP V NP 0.7 VP VP PP 0.3 Eglish 3
4 MLE Maximum likelihood estimatio picks the values for the model parameters that maximize the likelihood of the traiig data MLE Maximum likelihood estimatio picks the values for the model parameters that maximize the likelihood of the traiig data parameters S NP VP 0.9 S VP 0.1 NP Det A N 0.5 NP NP PP 0.3 NP PropN 0.2 A ε 0.6 A Adj A 0.4 PP Prep NP 1.0 VP V NP 0.7 VP VP PP 0.3 model ( parameter values parameters S NP VP S VP NP Det A N NP NP PP NP PropN A ε A Adj A PP Prep NP VP V NP VP VP PP model ( parameter values MLE(data = argmax p (data = argmax # p (data i i = argmax log( # p (data i i If this is what you wat to optimize, you ca do NO BETTER tha MLE MLE example You flip a coi 100 times. 60 times you get heads. What is the MLE for heads? p(head = 0.60 What is the likelihood of the data uder this model (each coi flip is a data poit? likelihood(data = # i p (data i log( * = MLE Example Ca we do ay better? likelihood(data = # i p (data i p(heads = 0.5 log( * =-69.3 p(heads = 0.7 log( * =
5 Traiig logistic regressio models Idea 1: miimize the squared error (like liear regressio log P(1 1 P(1 = w 0 + w 1 + w w m x m # error(h = (y i h( f i 2 We do t kow what the actual probability values are Ideas? Traiig logistic regressio models Idea 2: maximum likelihood traiig MLE(data = argmax p (data = argmax w = argmax w p w (label i f i log p w (label i f i How do we solve this? Traiig logistic regressio models Idea 2: maximum likelihood traiig Covex fuctios Covex fuctios look somethig like: MLE(data = argmax p (data = argmax w = argmax w p w (label i f i log p w (label i f i Ufortuately, o closed form solutio. 1. plug i our logistic equatio 2. take partial derivatives ad solve What are some ice properties about covex fuctios? How ca we fid the miimum/maximum of a covex fuctio? 5
6 Fidig the miimum Oe approach: gradiet descet Partial derivatives give us the slope i that dimesio You re blidfolded, but you ca see out of the bottom of the blidfold to the groud right by your feet. I drop you off somewhere ad tell you that you re i a covex shaped valley ad escape is at the bottom/miimum. How do you get out? Approach: pick a startig poit (w repeat util likelihood ca t icrease i ay dimesio: pick a dimesio move a small amout i that dimesio towards icreasig likelihood (usig the derivative Gradiet descet pick a startig poit (w repeat util loss does t decrease i all dimesios: pick a dimesio move a small amout i that dimesio towards decreasig loss (usig the derivative w i = w i # d error(w dw i learig rate (how much we wat to move i the error directio Solvig covex fuctios Gradiet descet is just oe approach A whole field called covex optimizatio Lots of well kow methods Cojugate gradiet Geeralized Iterative Scalig (GIS Improved Iterative Scalig (IIS Limited-memory quasi-newto (L-BFGS The key: if we get a error fuctio that is covex, we ca miimize/maximize it (evetually 6
7 Aother thought experimet Aother thought experimet What is a 100,000-dimesioal space like? You re a 1-D creature, ad you decide to buy a 2-uit apartmet What is a 100,000-dimesioal space like? Your job s goig well ad you re makig good moey. You upgrade to a 2-D apartmet with 2-uits per dimesio 2 rooms (very, skiy rooms 4 rooms (very, flat rooms Aother thought experimet Aother thought experimet What is a 100,000-dimesioal space like? You get promoted agai ad start havig kids ad decide to upgrade to aother dimesio. What is a 100,000-dimesioal space like? Larry Page steps dow as CEO of google ad they ask you if you d like the job. You decide to upgrade to a 100,000 dimesioal apartmet. 8 rooms (very, ormal rooms Each time you add a dimesio, the amout of space you have to work with goes up expoetially How much room do you have? Ca you have a big party? 2 100,000 rooms (it s very quiet ad loely = ~10 30 rooms per perso if you ivited everyoe o the plaet 7
8 The challege Overfittig Because logistic regressio has fewer costraits (tha, say NB it has a lot more optios We re tryig to fid 100,000 w values (or a poit i a 100,000 dimesioal space It s easy for logistic regressio to fit to uaces i the data: overfittig Give these poits as traiig data, which is a better lie to lear to separate the poits? Prevetig overfittig Prevetig overfittig log P(1 1 P(1 = w 0 + w 1 + w w m x m log P(1 1 P(1 = w 0 + w 1 + w w m x m We wat to avoid ay sigle feature from havig too much weight We wat to avoid ay sigle feature from havig too much weight MLE(data = argmax w log p w (y f ormal MLE MLE(data = argmax w log p w (y f ormal MLE ideas? MLE(data = argmax w m log p w (y f 2 # $ w j j =1 regularized MLE 8
9 Prevetig overfittig: regularizatio NB vs. Logistic regressio MLE(data = argmax w m log p w (y f 2 # $ w regularized MLE j What affect will this have o the leared weights assumig a positive? pealize large weights ecourage smaller weights - still a covex problem - equivalet to assumig your w j are distributed from a Gaussia with mea 0 (called a prior j =1 NB ad logistic regressio look very similar both are probabilistic models both are liear both lear parameters that maximize the log-likelihood of the traiig data How are they differet? NB vs. Logistic regressio Some historical perspective NB Logistic regressio f 1 log(p( f 1 l + f 1 log(1 P( f 1 l log(p(l e w 0 +w 1 +w w mx m 1+ e w 0 +w 1 +w w mx m Estimates the weights uder the strict assumptio that the features are idepedet If NB assumptio does t hold, we ca adjust the weights to compesate for this Naïve bayes is called a geerative model; it models the joit distributio p(features, labels Logistic regressio is called a discrimiative model; it models the coditioal distributio directly p(labels features 9
10 Estimatig the best chess state Old school optimizatio Possible parses (or whatever have scores Pick the oe with the best score How do you defie the score? Completely ad hoc Throw aythig you wat ito the mix Add a bous for this, a pealty for that, etc. State evaluatio fuctio for chess Write a fuctio that takes as iput a state represetatio of tic tac toe ad scores how good it is for you if you re X. How would you do it? (Called a state evaluatio fuctio Old school optimizatio Learig adjust bouses ad pealties by had to improve performace. J Total kludge, but totally flexible too Ca throw i ay ituitios you might have But we re purists we oly use probabilities New revolutio? Probabilities 10
11 New revolutio? Exposé at 9 Probabilities Probabilistic Revolutio Not Really a Revolutio, Critics Say Probabilities o more tha scores i disguise We re just addig stuff up like the old corrupt regime did, admits spokesperso 83% of Probabilists Rally Behid Paradigm ^.2,.4,.6,.8 We re ot goa take your bait 1. Ca estimate our parameters automatically e.g., p(t7 t5, t6 (trigram probability from supervised or usupervised data 2. Our results are more meaigful Ca use probabilities to place bets, quatify risk e.g., how sure are we that this is the correct parse? 3. Our results ca be meaigfully combied modularity Multiply idep. coditioal probs ormalized, ulike scores p(eglish text * p(eglish phoemes Eglish text * p(jap. phoemes Eglish phoemes * p(jap. text Jap. phoemes p(sematics * p(sytax sematics * p(morphology sytax * p (phoology morphology * p(souds phoology Probabilists Regret Beig Boud by Priciple Probabilists Regret Beig Boud by Priciple Ad-hoc approach does have oe advatage Cosider e.g. Naïve Bayes for spam categorizatio: Buy this supercalifragilistic Gisu kife set for oly $39 today Some useful features: Cotais Buy Cotais supercalifragilistic Cotais a dollar amout uder $100 Cotais a imperative setece Readig level = 8 th grade Metios moey (use word classes ad/or regexp to detect this Ay problem with these features for NB? Buy this supercalifragilistic Gisu kife set for oly $39 today Naïve Bayes Cotais a dollar amout uder $100 Metios moey (use word classes ad/or regexp to detect this Spam < $ Moey amout ot-spam How likely is it to see both features i either class usig NB? Is this right? 11
12 Probabilists Regret Beig Boud by Priciple Buy this supercalifragilistic Gisu kife set for oly $39 today Naïve Bayes Cotais a dollar amout uder $100 Metios moey (use word classes ad/or regexp to detect this Spam < $ Moey amout ot-spam 0.5*0.9= *0.1=0.002 Overestimates The problem is that the features are ot idepedet NB vs. Logistic regressio Logistic regressio allows us to put i features that overlap ad adjust the probabilities accordigly Which to use? NB is better for small data sets: strog model assumptios keep the model from overfittig Logistic regressio is better for larger data sets: ca exploit the fact that NB assumptio is rarely true NB vs. Logistic regressio NB vs. Logistic regressio 12
13 Logistic regressio with more classes Challege: probabilistic modelig NB works o multiple classes Logistic regressio oly works o two classes Idea: somethig like logistic regressio, but with more classes Like NB, oe model per each class The model is a weight vector P(class 1 = e w 1,0 +w 1,1 +w 1, w 1,mx m P(class 2 = e w 2,0 +w 2,1 +w 2, w 2,mx m P(class 1 = e w 1,0 +w 1,1 +w 1, w 1,mx m P(class 2 = e w 2,0 +w 2,1 +w 2, w 2,mx m P(class 3 = e w 3,0 +w 3,1 +w 3, w 3,mx m These are supposed to be probabilities P(class 3 = e w 3,0 +w 3,1 +w 3, w 3,m x m P(class 1 + P(class 2 + P(class aythig wrog with this? Ideas? Maximum Etropy Modelig aka Multiomial Logistic Regressio Log-liear model Normalize each class probability by the sum over all the classes e w 1,0 +w 1,1 +w 1, w 1,mx m P(class 1 = P(class 1 + P(class 2 + P(class P(class 1 = C ew 1,0 +w 1,1 +w 1, w 1,m x m P(class i P(class 1 = C ew 1,0 +w 1,1 +w 1, w 1,mx m P(class i = ew 1,0 +w 1,1 +w 1, w 1,mx m C e w i,0 +w i,1 +w i, w i,m x m ormalizig costat $ C ' logp(class 1 = w 1,0 + w 1,1 + w 1, w 1,m x m log& # P(class i % ( - still just a liear combiatio of feature weightigs - class specific features 13
14 Traiig the model Ca still use maximum likelihood traiig MLE(data = argmax Use regularizatio # log p(label i f i MLE(data = argmax # log p(label i f i $ %R( Plug ito a covex optimizatio package there are a few complicatios, but this is the basic idea Maximum Etropy Suppose there are 10 classes, A through J. I do t give you ay other iformatio. Questio: Give a ew example m: what is your guess for p(c m? Suppose I tell you that 55% of all examples are i class A. Questio: Now what is your guess for p(c m? Suppose I also tell you that 10% of all examples cotai Buy ad 80% of these are i class A or C. Questio: Now what is your guess for p(c m, if m cotais Buy? Maximum Etropy Maximum Etropy A B C D E F G H I J prob Qualitatively Maximum etropy priciple: give the costraits, pick the probabilities as equally as possible Quatitatively Maximum etropy: give the costraits, pick the probabilities so as to maximize the etropy Etropy(model = c p(clog p(c A B C D E F G H I J prob Qualitatively Maximum etropy priciple: give the costraits, pick the probabilities as equally as possible Quatitatively Maximum etropy: give the costraits, pick the probabilities so as to maximize the etropy Etropy(model = c p(clog p(c 14
15 Maximum Etropy Maximum Etropy A B C D E F G H I J Buy Other Colum A sums to 0.55 ( 55% of all messages are i class A A B C D E F G H I J Buy Other Colum A sums to 0.55 Row Buy sums to 0.1 ( 10% of all messages cotai Buy Maximum Etropy Maximum Etropy A B C D E F G H I J Buy Other Colum A sums to 0.55 Row Buy sums to 0.1 (Buy, A ad (Buy, C cells sum to 0.08 ( 80% of the 10% Give these costraits, fill i cells as equally as possible : maximize the etropy (related to cross-etropy, perplexity Etropy = log log log Largest if probabilities are evely distributed A B C D E F G H I J Buy Other Colum A sums to 0.55 Row Buy sums to 0.1 (Buy, A ad (Buy, C cells sum to 0.08 ( 80% of the 10% Give these costraits, fill i cells as equally as possible : maximize the etropy Now p(buy, C =.029 ad p(c Buy =.29 We got a compromise: p(c Buy < p(a Buy <.55 15
16 Geeralizig to More Features What we just did Other <$100 A B C D E F G H Buy Other For each feature ( cotais Buy, see what fractio of traiig data has it May distributios p(c,m would predict these fractios Of these, pick distributio that has max etropy Amazig Theorem: The maximum etropy model is the same as the maximum likelihood model If we calculate the maximum likelihood parameters, we re also calculatig the maximum etropy model What to take home May learig approaches Bayesia approaches (of which NB is just oe Liear regressio Logistic regressio Maximum Etropy (multiomial logistic regressio SVMs Decisio trees Differet models have differet stregths/weakesses/uses Uderstad what the model is doig Uderstad what assumptios the model is makig Pick the model that makes the most sese for your problem/data Feature selectio is importat Articles discussio 16
I. Chi-squared Distributions
1 M 358K Supplemet to Chapter 23: CHI-SQUARED DISTRIBUTIONS, T-DISTRIBUTIONS, AND DEGREES OF FREEDOM To uderstad t-distributios, we first eed to look at aother family of distributios, the chi-squared distributios.
More informationModified Line Search Method for Global Optimization
Modified Lie Search Method for Global Optimizatio Cria Grosa ad Ajith Abraham Ceter of Excellece for Quatifiable Quality of Service Norwegia Uiversity of Sciece ad Techology Trodheim, Norway {cria, ajith}@q2s.tu.o
More informationReview: Classification Outline
Data Miig CS 341, Sprig 2007 Decisio Trees Neural etworks Review: Lecture 6: Classificatio issues, regressio, bayesia classificatio Pretice Hall 2 Data Miig Core Techiques Classificatio Clusterig Associatio
More informationLog-Linear Models a.k.a. Logistic Regression, Maximum Entropy Models
Log-Linear Models a.k.a. Logistic Regression, Maximum Entropy Models Natural Language Processing CS 6120 Spring 2014 Northeastern University David Smith (some slides from Jason Eisner and Dan Klein) summary
More informationChapter 7 Methods of Finding Estimators
Chapter 7 for BST 695: Special Topics i Statistical Theory. Kui Zhag, 011 Chapter 7 Methods of Fidig Estimators Sectio 7.1 Itroductio Defiitio 7.1.1 A poit estimator is ay fuctio W( X) W( X1, X,, X ) of
More informationChapter 6: Variance, the law of large numbers and the Monte-Carlo method
Chapter 6: Variace, the law of large umbers ad the Mote-Carlo method Expected value, variace, ad Chebyshev iequality. If X is a radom variable recall that the expected value of X, E[X] is the average value
More informationChapter 5 Unit 1. IET 350 Engineering Economics. Learning Objectives Chapter 5. Learning Objectives Unit 1. Annual Amount and Gradient Functions
Chapter 5 Uit Aual Amout ad Gradiet Fuctios IET 350 Egieerig Ecoomics Learig Objectives Chapter 5 Upo completio of this chapter you should uderstad: Calculatig future values from aual amouts. Calculatig
More informationHypothesis testing. Null and alternative hypotheses
Hypothesis testig Aother importat use of samplig distributios is to test hypotheses about populatio parameters, e.g. mea, proportio, regressio coefficiets, etc. For example, it is possible to stipulate
More informationSoving Recurrence Relations
Sovig Recurrece Relatios Part 1. Homogeeous liear 2d degree relatios with costat coefficiets. Cosider the recurrece relatio ( ) T () + at ( 1) + bt ( 2) = 0 This is called a homogeeous liear 2d degree
More informationCase Study. Normal and t Distributions. Density Plot. Normal Distributions
Case Study Normal ad t Distributios Bret Halo ad Bret Larget Departmet of Statistics Uiversity of Wiscosi Madiso October 11 13, 2011 Case Study Body temperature varies withi idividuals over time (it ca
More informationCS103A Handout 23 Winter 2002 February 22, 2002 Solving Recurrence Relations
CS3A Hadout 3 Witer 00 February, 00 Solvig Recurrece Relatios Itroductio A wide variety of recurrece problems occur i models. Some of these recurrece relatios ca be solved usig iteratio or some other ad
More informationIn nite Sequences. Dr. Philippe B. Laval Kennesaw State University. October 9, 2008
I ite Sequeces Dr. Philippe B. Laval Keesaw State Uiversity October 9, 2008 Abstract This had out is a itroductio to i ite sequeces. mai de itios ad presets some elemetary results. It gives the I ite Sequeces
More informationRepeating Decimals are decimal numbers that have number(s) after the decimal point that repeat in a pattern.
5.5 Fractios ad Decimals Steps for Chagig a Fractio to a Decimal. Simplify the fractio, if possible. 2. Divide the umerator by the deomiator. d d Repeatig Decimals Repeatig Decimals are decimal umbers
More informationMath C067 Sampling Distributions
Math C067 Samplig Distributios Sample Mea ad Sample Proportio Richard Beigel Some time betwee April 16, 2007 ad April 16, 2007 Examples of Samplig A pollster may try to estimate the proportio of voters
More informationGCSE STATISTICS. 4) How to calculate the range: The difference between the biggest number and the smallest number.
GCSE STATISTICS You should kow: 1) How to draw a frequecy diagram: e.g. NUMBER TALLY FREQUENCY 1 3 5 ) How to draw a bar chart, a pictogram, ad a pie chart. 3) How to use averages: a) Mea - add up all
More information5: Introduction to Estimation
5: Itroductio to Estimatio Cotets Acroyms ad symbols... 1 Statistical iferece... Estimatig µ with cofidece... 3 Samplig distributio of the mea... 3 Cofidece Iterval for μ whe σ is kow before had... 4 Sample
More informationHere are a couple of warnings to my students who may be here to get a copy of what happened on a day that you missed.
This documet was writte ad copyrighted by Paul Dawkis. Use of this documet ad its olie versio is govered by the Terms ad Coditios of Use located at http://tutorial.math.lamar.edu/terms.asp. The olie versio
More informationWeek 3 Conditional probabilities, Bayes formula, WEEK 3 page 1 Expected value of a random variable
Week 3 Coditioal probabilities, Bayes formula, WEEK 3 page 1 Expected value of a radom variable We recall our discussio of 5 card poker hads. Example 13 : a) What is the probability of evet A that a 5
More informationCS103X: Discrete Structures Homework 4 Solutions
CS103X: Discrete Structures Homewor 4 Solutios Due February 22, 2008 Exercise 1 10 poits. Silico Valley questios: a How may possible six-figure salaries i whole dollar amouts are there that cotai at least
More informationDetermining the sample size
Determiig the sample size Oe of the most commo questios ay statisticia gets asked is How large a sample size do I eed? Researchers are ofte surprised to fid out that the aswer depeds o a umber of factors
More informationProperties of MLE: consistency, asymptotic normality. Fisher information.
Lecture 3 Properties of MLE: cosistecy, asymptotic ormality. Fisher iformatio. I this sectio we will try to uderstad why MLEs are good. Let us recall two facts from probability that we be used ofte throughout
More informationMaximum Likelihood Estimators.
Lecture 2 Maximum Likelihood Estimators. Matlab example. As a motivatio, let us look at oe Matlab example. Let us geerate a radom sample of size 00 from beta distributio Beta(5, 2). We will lear the defiitio
More information1. C. The formula for the confidence interval for a population mean is: x t, which was
s 1. C. The formula for the cofidece iterval for a populatio mea is: x t, which was based o the sample Mea. So, x is guarateed to be i the iterval you form.. D. Use the rule : p-value
More informationOutput Analysis (2, Chapters 10 &11 Law)
B. Maddah ENMG 6 Simulatio 05/0/07 Output Aalysis (, Chapters 10 &11 Law) Comparig alterative system cofiguratio Sice the output of a simulatio is radom, the comparig differet systems via simulatio should
More informationA gentle introduction to Expectation Maximization
A getle itroductio to Expectatio Maximizatio Mark Johso Brow Uiversity November 2009 1 / 15 Outlie What is Expectatio Maximizatio? Mixture models ad clusterig EM for setece topic modelig 2 / 15 Why Expectatio
More informationCHAPTER 3 THE TIME VALUE OF MONEY
CHAPTER 3 THE TIME VALUE OF MONEY OVERVIEW A dollar i the had today is worth more tha a dollar to be received i the future because, if you had it ow, you could ivest that dollar ad ear iterest. Of all
More informationUniversal coding for classes of sources
Coexios module: m46228 Uiversal codig for classes of sources Dever Greee This work is produced by The Coexios Project ad licesed uder the Creative Commos Attributio Licese We have discussed several parametric
More informationNormal Distribution.
Normal Distributio www.icrf.l Normal distributio I probability theory, the ormal or Gaussia distributio, is a cotiuous probability distributio that is ofte used as a first approimatio to describe realvalued
More informationLearning objectives. Duc K. Nguyen - Corporate Finance 21/10/2014
1 Lecture 3 Time Value of Moey ad Project Valuatio The timelie Three rules of time travels NPV of a stream of cash flows Perpetuities, auities ad other special cases Learig objectives 2 Uderstad the time-value
More informationLogistic Regression. Chapter 12. 12.1 Modeling Conditional Probabilities
Chapter 12 Logistic Regressio 12.1 Modelig Coditioal Probabilities So far, we either looked at estimatig the coditioal expectatios of cotiuous variables (as i regressio), or at estimatig distributios.
More informationOverview of some probability distributions.
Lecture Overview of some probability distributios. I this lecture we will review several commo distributios that will be used ofte throughtout the class. Each distributio is usually described by its probability
More informationZ-TEST / Z-STATISTIC: used to test hypotheses about. µ when the population standard deviation is unknown
Z-TEST / Z-STATISTIC: used to test hypotheses about µ whe the populatio stadard deviatio is kow ad populatio distributio is ormal or sample size is large T-TEST / T-STATISTIC: used to test hypotheses about
More informationUniversity of California, Los Angeles Department of Statistics. Distributions related to the normal distribution
Uiversity of Califoria, Los Ageles Departmet of Statistics Statistics 100B Istructor: Nicolas Christou Three importat distributios: Distributios related to the ormal distributio Chi-square (χ ) distributio.
More informationDiscrete Mathematics and Probability Theory Spring 2014 Anant Sahai Note 13
EECS 70 Discrete Mathematics ad Probability Theory Sprig 2014 Aat Sahai Note 13 Itroductio At this poit, we have see eough examples that it is worth just takig stock of our model of probability ad may
More informationNon-life insurance mathematics. Nils F. Haavardsson, University of Oslo and DNB Skadeforsikring
No-life isurace mathematics Nils F. Haavardsso, Uiversity of Oslo ad DNB Skadeforsikrig Mai issues so far Why does isurace work? How is risk premium defied ad why is it importat? How ca claim frequecy
More informationSimple Annuities Present Value.
Simple Auities Preset Value. OBJECTIVES (i) To uderstad the uderlyig priciple of a preset value auity. (ii) To use a CASIO CFX-9850GB PLUS to efficietly compute values associated with preset value auities.
More informationSolutions to Selected Problems In: Pattern Classification by Duda, Hart, Stork
Solutios to Selected Problems I: Patter Classificatio by Duda, Hart, Stork Joh L. Weatherwax February 4, 008 Problem Solutios Chapter Bayesia Decisio Theory Problem radomized rules Part a: Let Rx be the
More informationINVESTMENT PERFORMANCE COUNCIL (IPC)
INVESTMENT PEFOMANCE COUNCIL (IPC) INVITATION TO COMMENT: Global Ivestmet Performace Stadards (GIPS ) Guidace Statemet o Calculatio Methodology The Associatio for Ivestmet Maagemet ad esearch (AIM) seeks
More informationCenter, Spread, and Shape in Inference: Claims, Caveats, and Insights
Ceter, Spread, ad Shape i Iferece: Claims, Caveats, ad Isights Dr. Nacy Pfeig (Uiversity of Pittsburgh) AMATYC November 2008 Prelimiary Activities 1. I would like to produce a iterval estimate for the
More informationSECTION 1.5 : SUMMATION NOTATION + WORK WITH SEQUENCES
SECTION 1.5 : SUMMATION NOTATION + WORK WITH SEQUENCES Read Sectio 1.5 (pages 5 9) Overview I Sectio 1.5 we lear to work with summatio otatio ad formulas. We will also itroduce a brief overview of sequeces,
More informationNow here is the important step
LINEST i Excel The Excel spreadsheet fuctio "liest" is a complete liear least squares curve fittig routie that produces ucertaity estimates for the fit values. There are two ways to access the "liest"
More informationIncremental calculation of weighted mean and variance
Icremetal calculatio of weighted mea ad variace Toy Fich faf@cam.ac.uk dot@dotat.at Uiversity of Cambridge Computig Service February 009 Abstract I these otes I eplai how to derive formulae for umerically
More informationLecture 13. Lecturer: Jonathan Kelner Scribe: Jonathan Pines (2009)
18.409 A Algorithmist s Toolkit October 27, 2009 Lecture 13 Lecturer: Joatha Keler Scribe: Joatha Pies (2009) 1 Outlie Last time, we proved the Bru-Mikowski iequality for boxes. Today we ll go over the
More informationUC Berkeley Department of Electrical Engineering and Computer Science. EE 126: Probablity and Random Processes. Solutions 9 Spring 2006
Exam format UC Bereley Departmet of Electrical Egieerig ad Computer Sciece EE 6: Probablity ad Radom Processes Solutios 9 Sprig 006 The secod midterm will be held o Wedesday May 7; CHECK the fial exam
More informationTHE REGRESSION MODEL IN MATRIX FORM. For simple linear regression, meaning one predictor, the model is. for i = 1, 2, 3,, n
We will cosider the liear regressio model i matrix form. For simple liear regressio, meaig oe predictor, the model is i = + x i + ε i for i =,,,, This model icludes the assumptio that the ε i s are a sample
More informationAsymptotic Growth of Functions
CMPS Itroductio to Aalysis of Algorithms Fall 3 Asymptotic Growth of Fuctios We itroduce several types of asymptotic otatio which are used to compare the performace ad efficiecy of algorithms As we ll
More information0.7 0.6 0.2 0 0 96 96.5 97 97.5 98 98.5 99 99.5 100 100.5 96.5 97 97.5 98 98.5 99 99.5 100 100.5
Sectio 13 Kolmogorov-Smirov test. Suppose that we have a i.i.d. sample X 1,..., X with some ukow distributio P ad we would like to test the hypothesis that P is equal to a particular distributio P 0, i.e.
More informationThe analysis of the Cournot oligopoly model considering the subjective motive in the strategy selection
The aalysis of the Courot oligopoly model cosiderig the subjective motive i the strategy selectio Shigehito Furuyama Teruhisa Nakai Departmet of Systems Maagemet Egieerig Faculty of Egieerig Kasai Uiversity
More information.04. This means $1000 is multiplied by 1.02 five times, once for each of the remaining sixmonth
Questio 1: What is a ordiary auity? Let s look at a ordiary auity that is certai ad simple. By this, we mea a auity over a fixed term whose paymet period matches the iterest coversio period. Additioally,
More informationTotally Corrective Boosting Algorithms that Maximize the Margin
Mafred K. Warmuth mafred@cse.ucsc.edu Ju Liao liaoju@cse.ucsc.edu Uiversity of Califoria at Sata Cruz, Sata Cruz, CA 95064, USA Guar Rätsch Guar.Raetsch@tuebige.mpg.de Friedrich Miescher Laboratory of
More informationHypergeometric Distributions
7.4 Hypergeometric Distributios Whe choosig the startig lie-up for a game, a coach obviously has to choose a differet player for each positio. Similarly, whe a uio elects delegates for a covetio or you
More informationChapter 14 Nonparametric Statistics
Chapter 14 Noparametric Statistics A.K.A. distributio-free statistics! Does ot deped o the populatio fittig ay particular type of distributio (e.g, ormal). Sice these methods make fewer assumptios, they
More informationVladimir N. Burkov, Dmitri A. Novikov MODELS AND METHODS OF MULTIPROJECTS MANAGEMENT
Keywords: project maagemet, resource allocatio, etwork plaig Vladimir N Burkov, Dmitri A Novikov MODELS AND METHODS OF MULTIPROJECTS MANAGEMENT The paper deals with the problems of resource allocatio betwee
More informationINFINITE SERIES KEITH CONRAD
INFINITE SERIES KEITH CONRAD. Itroductio The two basic cocepts of calculus, differetiatio ad itegratio, are defied i terms of limits (Newto quotiets ad Riema sums). I additio to these is a third fudametal
More informationDomain 1: Designing a SQL Server Instance and a Database Solution
Maual SQL Server 2008 Desig, Optimize ad Maitai (70-450) 1-800-418-6789 Domai 1: Desigig a SQL Server Istace ad a Database Solutio Desigig for CPU, Memory ad Storage Capacity Requiremets Whe desigig a
More informationLecture 4: Cauchy sequences, Bolzano-Weierstrass, and the Squeeze theorem
Lecture 4: Cauchy sequeces, Bolzao-Weierstrass, ad the Squeeze theorem The purpose of this lecture is more modest tha the previous oes. It is to state certai coditios uder which we are guarateed that limits
More informationBENEFIT-COST ANALYSIS Financial and Economic Appraisal using Spreadsheets
BENEIT-CST ANALYSIS iacial ad Ecoomic Appraisal usig Spreadsheets Ch. 2: Ivestmet Appraisal - Priciples Harry Campbell & Richard Brow School of Ecoomics The Uiversity of Queeslad Review of basic cocepts
More informationListing terms of a finite sequence List all of the terms of each finite sequence. a) a n n 2 for 1 n 5 1 b) a n for 1 n 4 n 2
74 (4 ) Chapter 4 Sequeces ad Series 4. SEQUENCES I this sectio Defiitio Fidig a Formula for the th Term The word sequece is a familiar word. We may speak of a sequece of evets or say that somethig is
More informationSAMPLE QUESTIONS FOR FINAL EXAM. (1) (2) (3) (4) Find the following using the definition of the Riemann integral: (2x + 1)dx
SAMPLE QUESTIONS FOR FINAL EXAM REAL ANALYSIS I FALL 006 3 4 Fid the followig usig the defiitio of the Riema itegral: a 0 x + dx 3 Cosider the partitio P x 0 3, x 3 +, x 3 +,......, x 3 3 + 3 of the iterval
More informationOverview. Learning Objectives. Point Estimate. Estimation. Estimating the Value of a Parameter Using Confidence Intervals
Overview Estimatig the Value of a Parameter Usig Cofidece Itervals We apply the results about the sample mea the problem of estimatio Estimatio is the process of usig sample data estimate the value of
More informationYour organization has a Class B IP address of 166.144.0.0 Before you implement subnetting, the Network ID and Host ID are divided as follows:
Subettig Subettig is used to subdivide a sigle class of etwork i to multiple smaller etworks. Example: Your orgaizatio has a Class B IP address of 166.144.0.0 Before you implemet subettig, the Network
More information1 Correlation and Regression Analysis
1 Correlatio ad Regressio Aalysis I this sectio we will be ivestigatig the relatioship betwee two cotiuous variable, such as height ad weight, the cocetratio of a ijected drug ad heart rate, or the cosumptio
More informationTime Value of Money, NPV and IRR equation solving with the TI-86
Time Value of Moey NPV ad IRR Equatio Solvig with the TI-86 (may work with TI-85) (similar process works with TI-83, TI-83 Plus ad may work with TI-82) Time Value of Moey, NPV ad IRR equatio solvig with
More informationEstimating Probability Distributions by Observing Betting Practices
5th Iteratioal Symposium o Imprecise Probability: Theories ad Applicatios, Prague, Czech Republic, 007 Estimatig Probability Distributios by Observig Bettig Practices Dr C Lych Natioal Uiversity of Irelad,
More information1 Computing the Standard Deviation of Sample Means
Computig the Stadard Deviatio of Sample Meas Quality cotrol charts are based o sample meas ot o idividual values withi a sample. A sample is a group of items, which are cosidered all together for our aalysis.
More informationMann-Whitney U 2 Sample Test (a.k.a. Wilcoxon Rank Sum Test)
No-Parametric ivariate Statistics: Wilcoxo-Ma-Whitey 2 Sample Test 1 Ma-Whitey 2 Sample Test (a.k.a. Wilcoxo Rak Sum Test) The (Wilcoxo-) Ma-Whitey (WMW) test is the o-parametric equivalet of a pooled
More informationMathematical goals. Starting points. Materials required. Time needed
Level A1 of challege: C A1 Mathematical goals Startig poits Materials required Time eeded Iterpretig algebraic expressios To help learers to: traslate betwee words, symbols, tables, ad area represetatios
More informationThe following example will help us understand The Sampling Distribution of the Mean. C1 C2 C3 C4 C5 50 miles 84 miles 38 miles 120 miles 48 miles
The followig eample will help us uderstad The Samplig Distributio of the Mea Review: The populatio is the etire collectio of all idividuals or objects of iterest The sample is the portio of the populatio
More informationTime Value of Money. First some technical stuff. HP10B II users
Time Value of Moey Basis for the course Power of compoud iterest $3,600 each year ito a 401(k) pla yields $2,390,000 i 40 years First some techical stuff You will use your fiacial calculator i every sigle
More information5.4 Amortization. Question 1: How do you find the present value of an annuity? Question 2: How is a loan amortized?
5.4 Amortizatio Questio 1: How do you fid the preset value of a auity? Questio 2: How is a loa amortized? Questio 3: How do you make a amortizatio table? Oe of the most commo fiacial istrumets a perso
More informationLesson 17 Pearson s Correlation Coefficient
Outlie Measures of Relatioships Pearso s Correlatio Coefficiet (r) -types of data -scatter plots -measure of directio -measure of stregth Computatio -covariatio of X ad Y -uique variatio i X ad Y -measurig
More informationAP Calculus AB 2006 Scoring Guidelines Form B
AP Calculus AB 6 Scorig Guidelies Form B The College Board: Coectig Studets to College Success The College Board is a ot-for-profit membership associatio whose missio is to coect studets to college success
More informationTradigms of Astundithi and Toyota
Tradig the radomess - Desigig a optimal tradig strategy uder a drifted radom walk price model Yuao Wu Math 20 Project Paper Professor Zachary Hamaker Abstract: I this paper the author iteds to explore
More informationApplication and research of fuzzy clustering analysis algorithm under micro-lecture English teaching mode
SHS Web of Cofereces 25, shscof/20162501018 Applicatio ad research of fuzzy clusterig aalysis algorithm uder micro-lecture Eglish teachig mode Yig Shi, Wei Dog, Chuyi Lou & Ya Dig Qihuagdao Istitute of
More informationSection 11.3: The Integral Test
Sectio.3: The Itegral Test Most of the series we have looked at have either diverged or have coverged ad we have bee able to fid what they coverge to. I geeral however, the problem is much more difficult
More informationChapter 7: Confidence Interval and Sample Size
Chapter 7: Cofidece Iterval ad Sample Size Learig Objectives Upo successful completio of Chapter 7, you will be able to: Fid the cofidece iterval for the mea, proportio, ad variace. Determie the miimum
More information1 The Gaussian channel
ECE 77 Lecture 0 The Gaussia chael Objective: I this lecture we will lear about commuicatio over a chael of practical iterest, i which the trasmitted sigal is subjected to additive white Gaussia oise.
More informationPresent Value Factor To bring one dollar in the future back to present, one uses the Present Value Factor (PVF): Concept 9: Present Value
Cocept 9: Preset Value Is the value of a dollar received today the same as received a year from today? A dollar today is worth more tha a dollar tomorrow because of iflatio, opportuity cost, ad risk Brigig
More informationPlug-in martingales for testing exchangeability on-line
Plug-i martigales for testig exchageability o-lie Valetia Fedorova, Alex Gammerma, Ilia Nouretdiov, ad Vladimir Vovk Computer Learig Research Cetre Royal Holloway, Uiversity of Lodo, UK {valetia,ilia,alex,vovk}@cs.rhul.ac.uk
More informationPre-Suit Collection Strategies
Pre-Suit Collectio Strategies Writte by Charles PT Phoeix How to Decide Whether to Pursue Collectio Calculatig the Value of Collectio As with ay busiess litigatio, all factors associated with the process
More informationBetting on Football Pools
Bettig o Football Pools by Edward A. Beder I a pool, oe tries to guess the wiers i a set of games. For example, oe may have te matches this weeked ad oe bets o who the wiers will be. We ve put wiers i
More informationGroups of diverse problem solvers can outperform groups of high-ability problem solvers
Groups of diverse problem solvers ca outperform groups of high-ability problem solvers Lu Hog ad Scott E. Page Michiga Busiess School ad Complex Systems, Uiversity of Michiga, A Arbor, MI 48109-1234; ad
More informationBond Valuation I. What is a bond? Cash Flows of A Typical Bond. Bond Valuation. Coupon Rate and Current Yield. Cash Flows of A Typical Bond
What is a bod? Bod Valuatio I Bod is a I.O.U. Bod is a borrowig agreemet Bod issuers borrow moey from bod holders Bod is a fixed-icome security that typically pays periodic coupo paymets, ad a pricipal
More informationUncertainty Chapter 13. Mausam (Based on slides by UW-AI faculty)
Ucertait Chapter 3 Mausam Based o slides b UW-AI facult Kowledge Represetatio KR Laguage Otological Commitmet Epistemological Commitmet ropositioal Logic facts true false ukow First Order Logic facts objects
More informationPROCEEDINGS OF THE YEREVAN STATE UNIVERSITY AN ALTERNATIVE MODEL FOR BONUS-MALUS SYSTEM
PROCEEDINGS OF THE YEREVAN STATE UNIVERSITY Physical ad Mathematical Scieces 2015, 1, p. 15 19 M a t h e m a t i c s AN ALTERNATIVE MODEL FOR BONUS-MALUS SYSTEM A. G. GULYAN Chair of Actuarial Mathematics
More informationApproximating Area under a curve with rectangles. To find the area under a curve we approximate the area using rectangles and then use limits to find
1.8 Approximatig Area uder a curve with rectagles 1.6 To fid the area uder a curve we approximate the area usig rectagles ad the use limits to fid 1.4 the area. Example 1 Suppose we wat to estimate 1.
More informationConfidence Intervals. CI for a population mean (σ is known and n > 30 or the variable is normally distributed in the.
Cofidece Itervals A cofidece iterval is a iterval whose purpose is to estimate a parameter (a umber that could, i theory, be calculated from the populatio, if measuremets were available for the whole populatio).
More information3 Basic Definitions of Probability Theory
3 Basic Defiitios of Probability Theory 3defprob.tex: Feb 10, 2003 Classical probability Frequecy probability axiomatic probability Historical developemet: Classical Frequecy Axiomatic The Axiomatic defiitio
More informationMARTINGALES AND A BASIC APPLICATION
MARTINGALES AND A BASIC APPLICATION TURNER SMITH Abstract. This paper will develop the measure-theoretic approach to probability i order to preset the defiitio of martigales. From there we will apply this
More informationNATIONAL SENIOR CERTIFICATE GRADE 12
NATIONAL SENIOR CERTIFICATE GRADE MATHEMATICS P EXEMPLAR 04 MARKS: 50 TIME: 3 hours This questio paper cosists of 8 pages ad iformatio sheet. Please tur over Mathematics/P DBE/04 NSC Grade Eemplar INSTRUCTIONS
More informationInfinite Sequences and Series
CHAPTER 4 Ifiite Sequeces ad Series 4.1. Sequeces A sequece is a ifiite ordered list of umbers, for example the sequece of odd positive itegers: 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23, 25, 27, 29...
More informationLecture 3. denote the orthogonal complement of S k. Then. 1 x S k. n. 2 x T Ax = ( ) λ x. with x = 1, we have. i = λ k x 2 = λ k.
18.409 A Algorithmist s Toolkit September 17, 009 Lecture 3 Lecturer: Joatha Keler Scribe: Adre Wibisoo 1 Outlie Today s lecture covers three mai parts: Courat-Fischer formula ad Rayleigh quotiets The
More informationCutting-Plane Training of Structural SVMs
Cuttig-Plae Traiig of Structural SVMs Thorste Joachims, Thomas Filey, ad Chu-Nam Joh Yu Abstract Discrimiative traiig approaches like structural SVMs have show much promise for buildig highly complex ad
More informationwhere: T = number of years of cash flow in investment's life n = the year in which the cash flow X n i = IRR = the internal rate of return
EVALUATING ALTERNATIVE CAPITAL INVESTMENT PROGRAMS By Ke D. Duft, Extesio Ecoomist I the March 98 issue of this publicatio we reviewed the procedure by which a capital ivestmet project was assessed. The
More informationChair for Network Architectures and Services Institute of Informatics TU München Prof. Carle. Network Security. Chapter 2 Basics
Chair for Network Architectures ad Services Istitute of Iformatics TU Müche Prof. Carle Network Security Chapter 2 Basics 2.4 Radom Number Geeratio for Cryptographic Protocols Motivatio It is crucial to
More informationSequences and Series
CHAPTER 9 Sequeces ad Series 9.. Covergece: Defiitio ad Examples Sequeces The purpose of this chapter is to itroduce a particular way of geeratig algorithms for fidig the values of fuctios defied by their
More informationAP Calculus BC 2003 Scoring Guidelines Form B
AP Calculus BC Scorig Guidelies Form B The materials icluded i these files are iteded for use by AP teachers for course ad exam preparatio; permissio for ay other use must be sought from the Advaced Placemet
More informationChapter 7 - Sampling Distributions. 1 Introduction. What is statistics? It consist of three major areas:
Chapter 7 - Samplig Distributios 1 Itroductio What is statistics? It cosist of three major areas: Data Collectio: samplig plas ad experimetal desigs Descriptive Statistics: umerical ad graphical summaries
More informationHow To Solve The Homewor Problem Beautifully
Egieerig 33 eautiful Homewor et 3 of 7 Kuszmar roblem.5.5 large departmet store sells sport shirts i three sizes small, medium, ad large, three patters plaid, prit, ad stripe, ad two sleeve legths log
More informationFinding the circle that best fits a set of points
Fidig the circle that best fits a set of poits L. MAISONOBE October 5 th 007 Cotets 1 Itroductio Solvig the problem.1 Priciples............................... Iitializatio.............................
More information