01 03 Sentiment Analysis in Twitter

8/9/2019 01 03 Sentiment Analysis in Twitter

1/12

Sentiment Analysis in Twit

ContAyushi Dalmia (ayushi.Dalmia@researc

Mayank Gupta(mayank.g@studeArpit Kumar Jaiswal(arpitkumar.jaiswal@studen

Chinthala Tharun Reddy(tharun.chinthala@studen

Course: Information Retrieval and Extraction, IIIT HInstructor: Dr. Vasude

Source Code: [http://github.com/mayank93/Twitter-Sentiment-

Demo: [http://10.2.4.49:8000/TSAA/]


2/12

What is Sentiment Analys

It is classification of the polarity of agiven text in the document, sentenceor phrase

The goal is to determine whether theexpressed opinion in the text ispositive, negative or neutral.


3/12

Negative

Neutral


4/12

Why is Sentiment Analysis Impo

Microblogging has become popular communication tool

Opinion of the mass is important Political party may want to know whether people support their pro

not.

Before investing into a company, one can leverage the sentiment opeople for the company to find out where it stands.

A company might want find out the reviews of its products


5/12

Using Twitter for Sentiment An

Popular microblogging site Short Text Messages of 140 characters

240+ million active users

500 million tweets are generated everyday

Twitter audience varies from common man to celebriti

Users often discuss current affairs and share personal vvarious subjects

Tweets are small in length and hence unambiguous


6/12

Problem Statemen

The problem at hand consists of two subtasks:

Phrase Level Sentiment Analysis in Twitter :

Given a message containing a marked instance of a word or a pdetermine whether that instance is positive, negative or neutracontext.

Sentence Level Sentiment Analysis in Twitter:

Given a message, decide whether the message is of positive, neneutral sentiment. For messages conveying both a positive and sentiment, whichever is the stronger sentiment should be chos

The task is inspired from SemEval 2013 , Task 9 : Sentiment Analysis in Twitter


7/12

Challenges Tweets are highly unstructured and also non-grammatical

Out of Vocabulary Words

Lexical Variation

Extensive usage of acronyms like asap, lol, afaik


8/12

Approach

SVM Classifier and Prediction

Feature Extractor

Preprocessing

Tokeniser

Tweet Downloader


9/12

Tweet Downloader

Download the tweets using Twitter API Tokenisation

Twitter specific POS Tagger developed by ARK Social Media Search

Preprocessing

Removing non-English Tweets

Replacing Emoticons by their polarity Remove URL, Target Mentions, Hashtags, Numbers.

Replace Negative Mentions

Replace Sequence of Repeated Characters eg. coooooooool by c

Remove Nouns and Prepositions

Approach


10/12


11/12

ResultsA baseline model by taking the unigrams, bigrams and trigramcompare it with the feature based model for both the sub-tas

Sub-Task Baseline Model Feature Based

Model

Baseline + Feature Base

Model

Phrase Based 62.24 % 77.33% 79.90%

Sentence Based 52.54% 57.57% 58.36%

Accuracy

F1 Score

Sub-Task Baseline Model Feature BasedModel

Baseline + Feature BaseModel

Phrase Based 76.27* 75.23 75.98

Sentence Based 55.70 59.86 60.55

*Classifies in positive classes only, hence high recall.


12/12

Conclusion

We investigated two kinds of models: Baseline and Feature

Models and demonstrate that combination of both these mperform the best.

For our feature-based approach, feature analysis reveals thamost important features are those that combine the prior pwords and their parts-of-speech tags.

Documents

01 03 Sentiment Analysis in Twitter