[PDF] [PDF] Text Classification and Naïve Bayes - Stanford University

[PDF] Naive Bayes classifier

Example of Bayes Theorem • Given: – A doctor knows that Cold causes fever 50 of the time Example of Naïve Bayes Classifier P(Refund=YesNo) = 3/7

[PDF] Naive Bayes Classifiers

Properties of Bayes classifiers Naive Bayes classifiers Parameter estimation, properties, example Dealing with sparse data Application: email classification

[PDF] Naïve Bayes Classifier - UCR CS

We are about to see some of the mathematical formalisms, and more examples, but keep in mind the basic idea Find out the probability of the previously unseen  

[PDF] Text Classification and Naïve Bayes - Stanford University

Text Classification: definition • Input: • a document d Text Classification and Naïve Bayes Formalizing the Naïve Bayes Classifier training examples was

[PDF] BAYESIAN CLASSIFICATION - Stony Brook Computer Science

Example of Bayes Classification: https://github com/varunon9/naive-bayes- classifier https://www slideshare net/ashrafmath/naive-bayes-15644818 

[PDF] Naive-Bayes Classification Algorithm

Naive-Bayes Classification Algorithm 1 Introduction to Bayesian Classification The Bayesian Classification represents a supervised learning method as well as  

[PDF] Naïve Bayes Lecture 17 - peoplecsailmitedu

Naïve Bayes Lecture 17 David Sontag New York Bayesian Learning • Use Bayes' rule Your second learning algorithm: MLE for mean of a Gaussian

[PDF] Naive Bayes Classifier

Artificial Intelligence Naïve Bayesian classifier classifier? Do we have enough examples to learn a good model? classify all the unlabeled examples in D

[PDF] naïve Bayesian assumption - IHES

Introduction and the most basic concepts Fundamentals of AI Conditional independence, Naïve Bayes and Bayesian Networks 

[PDF] Case Study I: Naïve Bayesian spam filtering

Example Assume that we have the following set of email classified as spam We want to use a naive Bayes classifier to build a spam filter based on the words

[PDF] name a few tests for aldehyde and ketones

[PDF] name change dentpin

[PDF] name change publication in newspaper cost nyc

[PDF] name of international airport in paris

[PDF] name of international airport in paris france

[PDF] name reservation california

[PDF] name reservation manitoba

[PDF] name reservation meaning

[PDF] name reservation online

[PDF] name reservation request form georgia

[PDF] name the 5 parts of the 5th amendment

[PDF] name the attributes of img tag

[PDF] name three colligative properties of solutions

[PDF] named entity recognition

[PDF] names from the inheritance cycle

Isthisspam?

Whatisthesubjectofthisarticle?•AntogonistsandInhibitors•BloodSupply•Chemistry•DrugTherapy•Embryology•Epidemiology•...6MeSHSubjectCategoryHierarchy?MEDLINE Article

I love this movie! It's sweet,

but with satirical humor. The dialogue is great and the adventure scenes are fun...

It manages to be whimsical

and romantic while laughing at the conventions of the fairy tale genre. I would recommend it to just about anyone. I've seen it several times, and I'm always happy to see it again whenever I have a friend who hasn't seen it yet! it I the to and seen yet would whimsical times sweet satirical adventure genre fairy humor have great 6 5 4 3 3 2 1 1 1 1 1 1 1 1 1 1 1 1

I love this movie! It's sweet,

but with satirical humor. The dialogue is great and the adventure scenes are fun...

It manages to be whimsical

and romantic while laughing at the conventions of the fairy tale genre. I would recommend it to just about anyone. I've seen it several times, and I'm always happy to see it again whenever I have a friend who hasn't seen it yet! it I the to and seen yet would whimsical times sweet satirical adventure genre fairy humor have great 6 5 4 3 3 2 1 1 1 1 1 1 1 1 1 1 1 1

I love this movie! It's sweet,

but with satirical humor. The dialogue is great and the adventure scenes are fun...

It manages to be whimsical

and romantic while laughing at the conventions of the fairy tale genre. I would recommend it to just about anyone. I've seen it several times, and I'm always happy to see it again whenever I have a friend who hasn't seen it yet! it I the to and seen yet would whimsical times sweet satirical adventure genre fairy humor have great 6 5 4 3 3 2 1 1 1 1 1 1 1 1 1 1 1 1

NaïveBayesClassifier(I)cMAP=argmaxc∈CP(c|d)=argmaxc∈CP(d|c)P(c)P(d)=argmaxc∈CP(d|c)P(c)MAP is "maximum a posteriori" = most likely classBayes RuleDropping the denominator

NaïveBayesClassifier(II)cMAP=argmaxc∈CP(d|c)P(c)Document d represented as features x1..xn=argmaxc∈CP(x1,x2,...,xn|c)P(c)

NaïveBayesClassifier(IV)How often does this class occur?cMAP=argmaxc∈CP(x1,x2,...,xn|c)P(c)O(|X|n•|C|)parametersWe can just count the relative frequencies in a corpusCouldonlybeestimatedifavery,verylargenumberoftrainingexampleswasavailable.

Laplace(add-1)smoothingforNaïveBayesˆP(wi|c)=count(wi,c)+1count(w,c)+1()w∈V∑=count(wi,c)+1count(w,cw∈V∑)#$%%&'(( + VˆP(wi|c)=count(wi,c)count(w,c)()w∈V∑

MultinomialNaïveBayes:Learning•CalculateP(cj)terms•ForeachcjinCdodocsj←alldocswithclass=cjP(wk|cj)←nk+αn+α|Vocabulary|P(cj)←|docsj||total # documents|•CalculateP(wk|cj)terms•Textj←singledoccontainingalldocsj•ForeachwordwkinVocabularynk←#ofoccurrencesofwkinTextj•Fromtrainingcorpus,extractVocabulary

PR RP F 2 2 )1( 1 )1( 1 1

53•Most(over)useddataset,21,578docs(each90types,200toknens)•9603training,3299testarticles(ModApte/Lewissplit)•118categories•Anarticlecanbeinmorethanonecategory•Learn118binarycategorydistinctions•Averagedocument(withatleastonecategory)has1.24classes•Onlyabout10outof118categoriesarelargeCommon categories(#train, #test)Evaluation:ClassicReuters-21578DataSet•Earn (2877, 1087) •Acquisitions (1650, 179)•Money-fx(538, 179)•Grain (433, 149)•Crude (389, 189)•Trade (369,119)•Interest (347, 131)•Ship (197, 89)•Wheat (212, 71)•Corn (182, 56)Sec. 15.2.4

54ReutersTextCategorizationdataset(Reuters-21578)document 2-MAR-1987 16:51:43.42livestockhog CHICAGO, March 2 -The American Pork Congress kicks off tomorrow, March 3, in Indianapolis with 160 of the nations pork producers from 44 member states determining industry positions on a number of issues, according to the National Pork Producers Council, NPPC.Delegates to the three day Congress will be considering 26 resolutions concerning various issues, including the future direction of farm policy and the tax law as it applies to the agriculture sector. The delegates will also debate whether to endorse concepts of a national PRV (pseudorabiesvirus) control and eradication program, the NPPC said.A large trade show, in conjunction with the congress, will feature the latest in technology in all areas of the industry, the NPPC added. ReuterSec. 15.2.4

56PerclassevaluationmeasuresRecall:Fractionofdocsinclassiclassifiedcorrectly:Precision:Fractionofdocsassignedclassithatareactuallyaboutclassi:Accuracy:(1-errorrate)Fractionofdocsclassifiedcorrectly:ciii∑ciji∑j∑ciicjij∑ciicijj∑Sec. 15.2.4

57Micro-vs.Macro-Averaging•Ifwehavemorethanoneclass,howdowecombinemultipleperformancemeasuresintoonequantity?•Macroaveraging:Computeperformanceforeachclass,thenaverage.•Microaveraging:Collectdecisionsforallclasses,computecontingencytable,evaluate.Sec. 15.2.4

62TheRealWorld•Gee,I'mbuildingatextclassifierforreal,now!•WhatshouldIdo?Sec. 15.3.1

63Notrainingdata?ManuallywrittenrulesIf(wheatorgrain)andnot(wholeorbread)thenCategorizeasgrain•Needcarefulcrafting•Humantuningondevelopmentdata•Time-consuming:2daysperclassSec. 15.3.1

64Verylittledata?•UseNaïveBayes•NaïveBayesisa"high-bias"algorithm(NgandJordan2002NIPS)•Getmorelabeleddata•Findcleverwaystogethumanstolabeldataforyou•Trysemi-supervisedtrainingmethods:•Bootstrapping,EMoverunlabeleddocuments,...Sec. 15.3.1

65Areasonableamountofdata?•Perfectforallthecleverclassifiers•SVM•RegularizedLogisticRegression•Youcanevenuseuser-interpretabledecisiontrees•Usersliketohack•ManagementlikesquickfixesSec. 15.3.1

66Ahugeamountofdata?•Canachievehighaccuracy!•Atacost:•SVMs(traintime)orkNN(testtime)canbetooslow•Regularizedlogisticregressioncanbesomewhatbetter•SoNaïveBayescancomebackintoitsownagain!Sec. 15.3.1

67Accuracyasafunctionofdatasize•Withenoughdata•ClassifiermaynotmatterSec. 15.3.1BrillandBankoonspellingcorrection

70Howtotweakperformance•Domain-specificfeaturesandweights:veryimportantinrealperformance•Sometimesneedtocollapseterms:•Partnumbers,chemicalformulas,...•Butstemminggenerallydoesn'thelp•Upweighting:Countingawordasifitoccurredtwice:•titlewords(Cohen&Singer1996)•firstsentenceofeachparagraph(Murata,1999)•Insentencesthatcontaintitlewords(Koetal,2002)Sec. 15.3.2

quotesdbs_dbs12.pdfusesText_18