Scribd
Upload
167 views

Haramaya University: Department of Information Science

This document outlines a proposed MSc thesis on opinion mining of Afaan Oromoo political texts. The study aims to construct a classification model to detect sentiment orientation in Afaan Oromoo texts. It will review literature on opinion mining, prepare a dataset of Afaan Oromoo opinions, select algorithms, develop an opinion mining framework, and evaluate performance. The significance is that it can help analyze public opinion, aid decision making, and further natural language processing for Afaan Oromoo and other local languages.

Uploaded by

Tofik Ahmed
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
167 views

Haramaya University: Department of Information Science

This document outlines a proposed MSc thesis on opinion mining of Afaan Oromoo political texts. The study aims to construct a classification model to detect sentiment orientation in Afaan Oromoo texts. It will review literature on opinion mining, prepare a dataset of Afaan Oromoo opinions, select algorithms, develop an opinion mining framework, and evaluate performance. The significance is that it can help analyze public opinion, aid decision making, and further natural language processing for Afaan Oromoo and other local languages.

Uploaded by

Tofik Ahmed
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 19

Haramaya University

POST GRADUATE PROGRAM DIRECTORATE

Department of Information Science


MSc Thesis Proposal

By:
Jemal Abate
1
OPINION MINING
FOR AFAAN
OROMOO
POLITICAL
2
1. Introduction
• Virmani and Tyagi (2014) defines that:
– Opinion is a person’s perspective about an issue or
an object
– Mining is extraction of knowledge from raw data,
– Opinion Mining (OM) is the technique used to
extract intelligent information based on a person’s
opinion from raw data available.
• According to Liu, (2012) OM is the field of
study that analyzes people’s opinions,
evaluations, attitudes, and emotions towards
products, services, organizations, and or issues,
3
1.1. Background of the Study
• Asking or giving an opinion to someone we
know is something that most people do in their
lives. In the past, when an individual needed
opinions, he/she asked friends and family
(Bakhtawar and Farouque, 2012).
• Nowadays no one is limited to asking about
product, issues or services, because there are
many user reviews and discussions in public
forums, social media, and blogs (Tullu, 2013).
4
Background Contd…

• According to Dhekra and Slim (2014) the high


rapidity and ease of use of the Internet &
social media there is:
– Various political information and opinions are
constantly published on social media.
– Most of political decisions in a much stronger
manner influenced by people’s who use social
media.
– Most political actors can evaluate their actual
standing using peoples opinion on social
media. 5
Backgroud Contd…

• According to Pang and Lillian, (2008) goods


& service consumption is not the only
motivation for people’s to express opinions
online rather a need for political information.
• According to Efron et al (2008) Opinion has a
great deal in politics, such as to:
– Understanding what voters are thinking,
– What public figures support or oppose,
– Improve the quality of information that voters
have access to,
6
1.2. Statement of the problem
• According to Bing (2012) opinions are key
influencers of our activities. We need to know
others’ opinions to make decision, Individual
consumers also want to know opinions of
existing users of a product or service before
purchasing it.
• Due to the availability of large volume of
people/customers opinion on social media
opinion mining can still be a difficult task
(Bakhtawar and Farouque, 2012).
7
Statement of the Problem Contd…

• The democratization of web publishing has led


to the explosion of the number of opinions
expressed over the internet, at the same time;
citizens are becoming more actively engaged
in policy issues with political organizations
and the government (Michael et al, 2011).
• As a result, the absence of automatic opinion
detection system in online content become
problem for identification of emerging societal
trends and analysis of public reactions to
policies. 8
Statement of the Problem Contd…

• Nowadays on social media peoples express


their feeling by using different languages
(Talvensaari et al., 2007).
• Afaan Oromoo is one of those languages in
which peoples express their feelings.
• The Afaan Oromoo language is a Cushitic
language spoken by about 40 million people in
Ethiopia. It is the third largest language in
Africa, spoken in Ethiopian & neighboring
countries like Kenya, Somalia and Djibouti
(Kualo, 2010). 9
Statement of the Problem Contd…

• As a result of the availability of large volume of


feelings expressed in Afaan Oromoo on social
media:
– Difficult to explore peoples opinion that were
expressed in Afaan Oromoo,
– Navigating to opinions page and monitoring them
on the web can be tiresome,
– Its time consuming task for a human to read them
one by one,
– Difficult to summarize them and organize them
into usable forms or to decide something from such
unstructured text. 10
Statement of the Problem Contd…

• The aim of this study is to construct a


classification model using opinion mining
techniques for Afaan Oromoo political
sentiments
• To this end, this study tries to answer the
following questions.
– How to prepare quality data set for experimentation?
– Which algorithm of opinion mining works better for
Afaan Oromoo?
– How much the proposed algorithm performs in
detecting opinion orientation?
11
1.3. Scope of the Study
• The study will only deal with mining of
opinions from sentiments containing
Afaan Oromoo texts, excluding the
mining of opinion in other types of
language andformat (e.g. Image, video,
gestures, etc).
• Sentiments collected will be analyzed and
summarized at document level analysis.

12
1.4. Significance of the study
• This study will have a significance to:
– Save the time of users while analysing mass
people`s opinion,
– Reduce the money and labour spent to find
peoples’s opinions,
– Make the right decision for governments or
organizations,
– Initiate further research in the area of natural
language processing for Afaan Oromoo
language as well as for other local languages.
13
1.5. Objectives of the Study
1.5.1. General Objective
• The general objective of this study will be to construct a classification
model using opinion mining techniques for Afaan Oromoo political
sentiments.
1.5.2. Specific Objectives
• To review related literature concerning opinion mining in local
and foreign languages,
• To prepare Afaan Oromoo opinionated texts for feature selection,
• To select the best algorithms and techniques that have been used
in opinion mining,
• To develop a prototype of opinion mining as a framework that
will serve as a model for Afaan Oromoo political opinion mining,
• To evaluate the performance of the selected opinion mining
approaches, 14
2. Methodology of the Study
2.1. Research Design
• According to Mettler (2014) experimental
research is a well-known method and well
suited for Design Science Research.
• For this study Design Science Research
will be applied, because, it`s concerned with
the artificial intelligence, i.e. information
technology artifacts, can use experiments to
thoroughly evaluate design alternatives and
identify superior manifestations to bring
improvements.
15
2.2. Data Collection and Preparation Method
• Political sentiments expressed in Afaan
Oromoo will be collected from online
reviews to develop a corpus.
• The sentiments will include both the
opinion object and the feelings expressed.
• Interview will be conducted with language
experts
• Document from different sources will also
be analyzed to have better understanding
about language. 16
2.3. Implementation Tools and
Techniques
• XAMPP: It is a simple and easy to create a
local web server for testing(Kasia, 2013).
• Python: is object-oriented, high level &
interpreted programming language supports
multiple programming patterns (Masheet,
2011).
• Cassandra: is a scalable, high-performance
distributed database designed to handle large
amount of data (Anonymous, n.d.).
• HTML: It is easy to learn, and allow web
based user interface creation (Shannon, 2012).
17
2.3. Evaluation Methods

• Precision: is the ratio of correctly predicted


+ve observations to the total predicted +ve
observatisons.
• Recall: is the ratio of correctly predicted +ve
observations to the all observations in actual
class - yes.
• F-measure: is the weighted average of
Precision and Recall.
18
Thank You!
19

You might also like