Scribd
Upload
55 views

Sentiment Analysis in Arabic Language Using Machine Learning Iraqi Dialect Case Study

The document describes a study on sentiment analysis of Iraqi dialect movie reviews using machine learning techniques. It presents a dataset of 1189 movie reviews labeled as positive or negative sentiment. The reviews were preprocessed and vectorized using TF-IDF before being classified with models like Naive Bayes, SVM, KNN, and logistic regression. The results showed that the proposed hybrid classifiers achieved the best performance with over 80% accuracy in detecting sentiment polarity in the Iraqi dialect movie reviews.

Uploaded by

Hussein Nassrullah
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
55 views

Sentiment Analysis in Arabic Language Using Machine Learning Iraqi Dialect Case Study

The document describes a study on sentiment analysis of Iraqi dialect movie reviews using machine learning techniques. It presents a dataset of 1189 movie reviews labeled as positive or negative sentiment. The reviews were preprocessed and vectorized using TF-IDF before being classified with models like Naive Bayes, SVM, KNN, and logistic regression. The results showed that the proposed hybrid classifiers achieved the best performance with over 80% accuracy in detecting sentiment polarity in the Iraqi dialect movie reviews.

Uploaded by

Hussein Nassrullah
Copyright
© © All Rights Reserved
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 15

8th Engineering and 2nd International Conference for College of Engineering

University of Baghdad 24–25 November / 2021

Sentiment Analysis in Arabic Language using


Machine Learning: Iraqi Dialect Case Study

Dr. Wameedh N. Flayyih


Hussein A. Nasrullah
Mohammed A. Nasrullah

21/7/2021 1
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

11/7/2021 2
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

INTRODUCTION

 Sentiment analysis is an application of natural language processing.


 The rapid growth of using social network increases the importance of natural
language processing

 This work:
o applies machine learning approaches for sentiment analysis of the Iraqi
dialect.
o extracts sentiments associated with positive or negative polarities
o considering the movies review comments as a case study.

11/7/2021 2
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

11/7/2021 3
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

11/7/2021 3
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

Simple example of data classification

11/7/2021 3
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

Dataset

 1189 movies reviews on Cinemana application


 The dataset was manually extracted and cleaned.
 Almost reviews in Arabic language - Iraqi dialects.
 Each review was labeled as either positive or negative
 Dataset was split into 70% for training and 30% for testing with 10-
fold cross validation.

11/7/2021 2
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

11/7/2021 2
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

Preprocessing

11/7/2021 2
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

Text Vectorization

 TF-IDF is a typical method to convert text to numerical vector


 TF-IDF stands for “Term Frequency — Inverse Document
Frequency”
 It depends on the number of times the word appears in the sentence and
in the document,
 Term Frequency depends on the number of times the word appears in the
sentence (review).
 Document Frequency depends on the number of times the word appears in the
sentence (all dataset)

11/7/2021 2
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

Classification

 Multinomial Naïve Bayes (MNB)

 Support Vector Machine (SVM)

 k-Nearest Neighbors (KNN)

 Logistic Regression (LR)

 Proposed hybrid classifiers

11/7/2021 2
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

Results

11/7/2021 2
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

Results

11/7/2021 2
8th Engineering and 2nd International Conference for College of Engineering
University of Baghdad 24–25 November / 2021

Thank for listening

11/7/2021 2

You might also like