Automated data collection tools and mature database technology lead to tremendous amounts of data stored in databases, data warehouses and other information repositories. Why Data Mining? What Is Data Mining? Data Mining: On What Kind of Data? Data Classification What is Sentiment Classification? Importance of Sentiment classification Twitter for Sentiment Classification Problem Statement Goal of this Classifications Method to be used Conclusion

1. BY:
Ishtiyak Rahman Shishir
Dept of CSE
Varendra University
Twitter Sentiment Classification

2. Index
 Why Data Mining?
 What Is Data Mining?
 Data Mining: On What Kind of Data?
 Data Classification
 What is Sentiment Classification?
 Importance of Sentiment classification
 Twitter for Sentiment Classification
 Problem Statement
 Goal of this Classifications
 Method to be used
 Conclusion

3. Why Data Mining?
Data explosion problem
Automated data collection tools and mature database technology lead to
tremendous amounts of data stored in databases, data warehouses and other
information repositories
 We are drowning in data, but starving for knowledge!
 Solution: Data warehousing and data mining
– Data warehousing and on-line analytical processing
– Extraction of interesting knowledge (rules, regularities, patterns,
constraints) from data in large databases

4. What Is Data Mining?
Data mining (knowledge discovery in databases)
Extraction of interesting (non-trivial, implicit, previously
unknown and potentially useful) information or patterns
from data in large databases

5. Data Mining: On What Kind of Data?
 Relational databases
 Data warehouses
 Transactional databases
 Advanced DB and information repositories

6. Data Mining
Online Analytical
Processing
Discovery Driven Methods
SQL Query Tools
Description Prediction
Classification Regressions

7. Data Classification
Classification consists of assigning a class label to a set of
unclassified cases.
 Supervised Classification
The set of possible classes is known in advance.
 Unsupervised Classification
Set of possible classes is not known. After classification we can try to
assign a name to that class. Unsupervised classification is called
clustering.

8. What is Sentiment Classification?
The process of computationally identifying and categorizing
opinions expressed in a piece of text.
The goal is to determine whether the writer's attitude towards a particular topic,
product, etc., is positive, negative, or neutral.

9. Importance of Sentiment classification
 Adjust marketing strategy
 Measure ROI of your marketing campaign
 Develop product quality
 Improve customer service
 Crisis management
 Lead generation
 Sales Revenue

10. Using Twitter for Sentiment
Classification
Most Popular microblogging site
 Short Text Messages of 140 characters
 328 million active users
 500 million tweets are generated everyday
 Twitter audience varies from common man to celebrities
 Users often discuss current affairs and share personal views on various s
ubjects
 Tweets are small in length and hence unambiguous
Last updated: 8/12/17 Source:https://www.omnicoreagency.com/twitter-statistics

11. Problem Statement
The problem at hand consists of two subtasks
– Emoticon-Hashtag Level Sentiment Analysis
Given a message containing hashtags and emoticons instance of a word
or a phrase, determine whether that instance is positive, negative or neu
tral in that context.
– Sentence Level Sentiment Analysis
Given a message containing a sentence, a word or a phrase,
determine whether that instance is positive, negative or neutral in that c
ontext.

12. Goal of this Classifications
There are Two goals to be achived
 Large Scale Implementations for Sentiment Classification
 Time efficiency for Sentiment Classification

13. Method to be used
We develop two systems
 MapReduce
 Apache Spark Framework
The task is inspired from MDPI by Andreas Kanavos, 2016 , Task : Twitter Sentiment Classification

14. Method to be used
MapReduce
 The process of large datasets on a classification
 It consists of two main procedures
- Map and Reduce

15. Method to be used
Apache Spark Framework
 Apache Spark is an open source big data processing framework built around
speed, ease of use
 Comprehensive, unified framework
 100 times faster in memory and 10 times faster even when running on disk
 It let quickly write applications in Java, Scala, or Python

16. Conclusion
Data mining is the best way to find out necessary informations and data classification
Make it more valuable. Hopefully, for huge amount of data,MapReduce model
and Spark Framework will help to expand the scalability of data and reduce execution
time.

17. References
 Bingwei Liu, Erik Blasch, Yu Chen, Dan Shen and Genshe Chen “Scalable
Sentiment Classification for Big Data Analysis Using Na ̈ıve Bayes Classifier”
on 2013 IEEE International Conference on Big Data
 Roseline Antai “Sentiment Classification Using Summaries: A Comparative
Investigation of Lexical and Statistical Approaches” on 2014 6th Computer
Science and Electronic Engineering Conference (CEEC)
 R. Suresh Ramanujam Ph.D, J. Nivedha, J. Kokila “SENTIMENT ANALYSIS
USING BIG DATA” on 2015 INTERNATIONAL CONFERENCE ON
COMPUTATION OF POWER, ENERGY, INFORMATION AND COMMUNICATION
 Divya Sehgal l and Dr. Ambuj Kumar Agarwal2 “Sentiment Analysis of Big Data Applications using T
witter Data with the Help of HADOOP Framework”
 RAVI VATRAPU1,2, RAGHAVA RAO MUKKAMALA1, ABID HUSSAIN1, AND BENJAMIN FLESCH1 “Social Set Ana
lysis:A set theoritical approch of big data analysis” on April 28, 2015
at IEEE

18. References
 Pragya Tripathi, Santosh Kr Vishwakarma, Ajay Lala “Sentiment Analysis of English Tweets Using
RapidMiner” on 2015 International Conference on Computational Intelligence and Communication
Networks
 Lukas Povoda, Radim Burget, Malay Kishore Dutta “Sentiment Analysis Based on Support Vector
Machine and Big Data”
 Beiming Sun, Vincent TY Ng “Analyzing Sentimental Influence of Posts on Social Networks” 2014
IEEE
 LI Bing, Keith C.C. Chan “A Fuzzy Logic Approach for Opinion Mining on Large Scale Twitter Data”
on 2014 IEEE/ACM
 Andreas Kanavos 1,*, Nikolaos Nodarakis 1, Spyros Sioutas 2, Athanasios Tsakalidis 1,
Dimitrios Tsolis 3 and Giannis Tzimas 4 “Large Scale Implementations for Twitter Sentiment Cla
ssification” on 4 March 2017 at MDPI

19. Thank You