Machine Learning and Sentiment Classification in Ruby

.BDIJOF -FBSOJOH BOE 4FOUJNFOU $MBTTJGJDBUJPO JO 3VCZ by Matt Drozdzynski
@matid

.BDIJOF -FBSOJOH … or how to teach your computer to
do back flips for you.

4FOUJNFOU $MBTTJGJDBUJPO … or how to quantify people’s opinions.

#euruko is definitely the most amazing Ruby conference ever!

I’ve been to many dreadful conferences, but #euruko is certainly
not one of them.

Ruby is a true delight compared to how horrendous Java
can be.

d JO 3VCZ

%BUB (BUIFSJOH

-BOHVBHF "DDVSBDZ 0% 25% 50% 75% 100% 2007 English Spanish
German Italian Polish

"OOPUBUJPOT … or I have the tweets—now what?

%BUB $MFBOJOH … or how to separate wheat from the
chaff.

'FBUVSF 3FEVDUJPO … or Matt’s crash course in selective ignorance.

$MBTTJGJDBUJPO … and the ‘not so rocket’ science behind it
all.

/BJWF #BZFT Simple and robust Assumes independence of features Scalable!

require "ankusa" require "ankusa/memory_storage" storage = Ankusa::MemoryStorage.new classifier = Ankusa::NaiveBayesClassifier.new(storage)
training.each do |tweet| classifier.train tweet.sentiment, tweet.to_s end sentiment = classifier.classify tweet.to_s

.BYJNVN &OUSPQZ No independence assumptions Suffers from overfitting Substantially slower
than Naive Bayes

require "maxent_string_classifier" classifier = MaxentStringClassifier::Loader.train(Classifier.root + "max_ent" + "data") classification
= classifier.classify tweet.to_s

4VQQPSU 7FDUPS .BDIJOFT Non-probabilistic binary linear classifier Only directly applicable
to two-class problems “Works by constructing a set of hyperplanes in a high or infinite dimensional space”—what?

require "eluka" classifier = Eluka::Model.new training.each do |tweet| classifier.add(tweet.features, tweet.sentiment)
end classifier.build sentiment = classifier.classify tweet.features

$PODMVTJPOT … or is the whole thing worth the hassle?

2VFTUJPOT

@matid spkr8.com/t/7678 bit.ly/matid-dissertation bit.ly/matid-dissertation-pdf 5IBOLT

Machine Learning and Sentiment Classification i...

Machine Learning and Sentiment Classification in Ruby

Matt D.

Other Decks in Programming

Featured

Transcript

.BDIJOF -FBSOJOH BOE 4FOUJNFOU $MBTTJGJDBUJPO JO 3VCZ by Matt Drozdzynski

.BDIJOF -FBSOJOH … or how to teach your computer to

4FOUJNFOU $MBTTJGJDBUJPO … or how to quantify people’s opinions.

#euruko is definitely the most amazing Ruby conference ever!

I’ve been to many dreadful conferences, but #euruko is certainly

Ruby is a true delight compared to how horrendous Java

d JO 3VCZ

%BUB (BUIFSJOH

-BOHVBHF "DDVSBDZ 0% 25% 50% 75% 100% 2007 English Spanish

"OOPUBUJPOT … or I have the tweets—now what?

%BUB $MFBOJOH … or how to separate wheat from the

'FBUVSF 3FEVDUJPO … or Matt’s crash course in selective ignorance.

$MBTTJGJDBUJPO … and the ‘not so rocket’ science behind it

/BJWF #BZFT Simple and robust Assumes independence of features Scalable!

require "ankusa" require "ankusa/memory_storage" storage = Ankusa::MemoryStorage.new classifier = Ankusa::NaiveBayesClassifier.new(storage)

.BYJNVN &OUSPQZ No independence assumptions Suffers from overfitting Substantially slower

require "maxent_string_classifier" classifier = MaxentStringClassifier::Loader.train(Classifier.root + "max_ent" + "data") classification

4VQQPSU 7FDUPS .BDIJOFT Non-probabilistic binary linear classifier Only directly applicable

require "eluka" classifier = Eluka::Model.new training.each do |tweet| classifier.add(tweet.features, tweet.sentiment)

$PODMVTJPOT … or is the whole thing worth the hassle?

2VFTUJPOT

@matid spkr8.com/t/7678 bit.ly/matid-dissertation bit.ly/matid-dissertation-pdf 5IBOLT