الفهرس 
الغلافOnly 14 pages are availabe for public view
 |  | from | 97 |  |  |
| | | from | 97 | | |
Abstract
Research Problem and Relevance
Recently, with the significant use of the internet in everything in our life from
commercial transactions, marketing, e-learning, political, manufacturing,
entertainment, and many more of our daily treatments. The opinions of people are
now occupying important place in most if not all fields, especially with the
increasing the people’s ability to share their opinion in very easy manner any time
any place. Thus, the user’s opinion has the ability to make radical changes in
several fields in particular with the high speed of the internet to share and publish
these opinions. As the outliers could be found in different data formats, it also
could be found in the opinionated text. Outlier opinion means that opinion does not
comply with the general opinions (unusual pattern).
Research Questions
This research will be guided by a main research question, which has been
formulated as follows:
”How can help the proposed mining outlier opinion model in detection the
anomalies from an opinionated comment? ”
Research Objectives
This research aims to support decision makers in different application domains.
The provided support specially helps those domains which extremely depend in
their works on the opinionated comments and reviews which are expressed from
users or general people. Therefore, the research objective is to detect the outliers
from people’s opinionated text to avoid their impact on the sentiment analysis
process and what follows this from changes in decision making process.
Research Methodology
The research methodology (As shown in Figure1.2) can be illustrated in the
following points:
1) Reviewing the related literature work in the research areas of opinion mining
and anomaly detection.
2) Conducting a comparison study to highlight the points of strengths and
weaknesses of related research work.
3) Development of the proposed mining outlier detection model “MOOM”.
4) Applying the proposed MOOM on a benchmark dataset
5) Testing different performance factors of the proposed model through an
experimental study.
6) Analyzing the results to evaluate the efficiency of the proposed model.