SlideShare a Scribd company logo
1 of 20
ARABIC SENTIMENT ANALYSIS
O F F O O D DEL IVERY SERVICES
REVIEWS
DR. DHEYA MUSTAFA
FACULTY OF ENGINEERING
THE HASHEMITE UNIVERSITY
Oct 24, 2023 ISNCC’2023, Doha, Qatar 1
OUTLINE
•Introduction
•Motivation
•Goal
•Contribution
•Proposed Model
•Performance Evaluation
•Conclusion
2
INTRODUCTION
• Food delivery services (FDSs) have
introduced diversity to the increasing
demand for online food delivery
marketplaces.
• Mostly third-party marketplace apps for
global ordering and delivery, such as
UberEATS, Talabat (the Arabic word for
orders), and Menulog.
• Employ a cost intensive, aggregator
business model and are in charge of all
delivery logistics. 3
MOTIVATION
• The majority of businesses want to
effectively increase customer
satisfaction by using data to identify
areas for improvement.
–Mostly based on reviews from
customers.
• Sentiment analysis (SA) can
determine the customers’ opinion
based on the written content.
4
GOAL
• Apply SA to Arabic content to give an even
better service, driven by the feedback of
customers.
• Offer a quick and efficient monitoring
system to handle the needs clients on a
large scale at a low cost and maximum
profit.
5
METHODOLOGY
• gathering Arabic reviews regarding FDS and
evaluating feelings using popular ML methods.
• Created an Arabic dialects lexicon including
various dialects and their standard orientation
in Arabic.
• Emotions are identified and analyzed using an
existing emotion lexicon.
• experiments were carried out before and
after applying cleaning and preprocessing .
• Negative sentiments were analyzed to identify
the possible causes. 6
7
DATASET CLEANING
• The raw datasets contain many empty and repetitive reviews that
should be efficiently eliminated.
• Involves removing duplications, identifying emojis, characters,
and word replacements, spelling correction, and annotation.
• SA of dialectical Arabic is challenging.
• We constructed a dialects dictionary manually to translate
dialectical phrases into equivalent MSA in the context of FDS.
• Popular English words written in the Arabic alphabet in the
reviews are included in the dialects dictionary as well.
– Example, the word “nice” is written as ‘‫نايس‬ ’ and is equivalent
to the word ‘‫جميل‬ ’ in Arabic.
8
9
ARABIC REVIEWS
PREPROCESSING
• Clean review texts are process using common NLP
tasks.
• normalization : represent step characters that have
more than one form like ( ،‫أ‬
،‫إ‬
‫آ‬ ) in a unified form.
• Also removed any punctuations, diacritics, and tatweel.
• Stopword removal: reduce the size of the text by
removing unnecessary particles and pronouns.
• Tokenization: split the text into single units such as
sentences and then into single words.
10
ARABIC REVIEWS
PREPROCESSING
• Part of Speech (PoS) tagging: annotate words in a text
based on their type and their relationships to
neighboring and related words in that text.
– Based on the Stanford tagger [35].
• The last NLP task involved removing suffixes or
prefixes.
– To efficiently convert each word in its base form without
extensions (Stemming)
– Used the tashaphyne stemmer
11
SENTIMENT
CLASSIFICATION
• The final phase was to build and train the machine
learning (ML) classifier using a labeled subset of
the dataset (Talabat FDS ).
• We tested the model on unseen reviews and
categorized them into positive, negative, or
neutral classes.
• ML models: DT and SVM
– Most frequently used classifiers in Arabic SA
– Often provide the best results over other ML classifiers
EVALUATION
14
EVALUATION
15
Experiment
DT classifier SVM Classifier
P% R% F% Acc% P% R% F% Acc%
Before
cleaning
and pre-
processin
g
+ve 67 95 79 87 86 86
-ve 82 44 57 80 86 83
Neu 20 4 6 31 23 26
Avg 68 69 65 69 79 80 80 80
After
cleaning
and pre-
processin
g
+ve 72 91 81 87 88 88
-ve 76 60 67 77 89 83
Neu 20 3 5 33 10 15
Avg 69 73 69 73 79 82 80 82
+ve: positive -ve: negative Neu: neutral Avg:
average
EVALUATION
16
EVALUATION
17
Fig. 3. A confusion matrix of the SVM classifier after applying cleaning
and preprocessing phases.
EVALUATION
18
CONCLUSION
• Proposed a sentiment analysis model of Arabic dialect
reviews about FDS.
• applied intensive preprocessing and several aspects of
analyzing Arabic text to generate a dataset.
• examined two well-known ML classifiers over this
dataset and provided a comparative performance
evaluation
• Analyzed the most representative negative reviews
and produced a word cloud.
– The long waiting time, cold food, wrong order, and
bill were the most important factors contributing
to the negative sentiments.
19
Thank you
Oct. 23, 2023 ISNCC'2014, Doha, Qatar 20

More Related Content

Similar to ISNCC '23 Presentation.pptx

Serge astm-presentation-chicago-2014-final
Serge astm-presentation-chicago-2014-finalSerge astm-presentation-chicago-2014-final
Serge astm-presentation-chicago-2014-finalSerge Gladkoff
 
ETSC Job descriptions
ETSC Job descriptionsETSC Job descriptions
ETSC Job descriptionsPritam Dutta
 
Lead scoring case study presentation
Lead scoring case study presentationLead scoring case study presentation
Lead scoring case study presentationMithul Murugaadev
 
TAUS 2.0 and the Game Changers in Localization, by Jaap van der Meer, directo...
TAUS 2.0 and the Game Changers in Localization, by Jaap van der Meer, directo...TAUS 2.0 and the Game Changers in Localization, by Jaap van der Meer, directo...
TAUS 2.0 and the Game Changers in Localization, by Jaap van der Meer, directo...TAUS - The Language Data Network
 
Translation and localization process optimization - www.konsul.info
Translation and localization process optimization - www.konsul.infoTranslation and localization process optimization - www.konsul.info
Translation and localization process optimization - www.konsul.infoDamian Pajnkiher
 
Building Large Arabic Multi-Domain Resources for Sentiment Analysis
Building Large Arabic Multi-Domain Resources for Sentiment Analysis Building Large Arabic Multi-Domain Resources for Sentiment Analysis
Building Large Arabic Multi-Domain Resources for Sentiment Analysis Hady Elsahar
 
Rapid Strategic SRE Assessments
Rapid Strategic SRE AssessmentsRapid Strategic SRE Assessments
Rapid Strategic SRE AssessmentsMarc Hornbeek
 
Ncerc rlmca202 adm m4 ssm
Ncerc rlmca202 adm m4 ssmNcerc rlmca202 adm m4 ssm
Ncerc rlmca202 adm m4 ssmssmarar
 
Classification of webpages as Ephemeral or Evergreen
Classification of webpages as Ephemeral or EvergreenClassification of webpages as Ephemeral or Evergreen
Classification of webpages as Ephemeral or EvergreenMonis Javed
 
North Star Case Study.pptx
North Star Case Study.pptxNorth Star Case Study.pptx
North Star Case Study.pptxHansaJoshi2
 
TCS Interview Questions and Answers 2022 | How to Crack TCS Interview for Fre...
TCS Interview Questions and Answers 2022 | How to Crack TCS Interview for Fre...TCS Interview Questions and Answers 2022 | How to Crack TCS Interview for Fre...
TCS Interview Questions and Answers 2022 | How to Crack TCS Interview for Fre...Simplilearn
 
Applying a Quality Framework to the In-depth Intrerview Method
Applying a Quality Framework to the In-depth Intrerview MethodApplying a Quality Framework to the In-depth Intrerview Method
Applying a Quality Framework to the In-depth Intrerview MethodRoller Research
 
The Quest for Quality at Speed
The Quest for Quality at SpeedThe Quest for Quality at Speed
The Quest for Quality at SpeedMarc Hornbeek
 

Similar to ISNCC '23 Presentation.pptx (20)

Arnab_Resume
Arnab_ResumeArnab_Resume
Arnab_Resume
 
ATA Certifications examination overview
ATA Certifications examination overviewATA Certifications examination overview
ATA Certifications examination overview
 
Quality and Localization Effectiveness
Quality and Localization EffectivenessQuality and Localization Effectiveness
Quality and Localization Effectiveness
 
Serge astm-presentation-chicago-2014-final
Serge astm-presentation-chicago-2014-finalSerge astm-presentation-chicago-2014-final
Serge astm-presentation-chicago-2014-final
 
ETSC Job descriptions
ETSC Job descriptionsETSC Job descriptions
ETSC Job descriptions
 
Lead scoring case study presentation
Lead scoring case study presentationLead scoring case study presentation
Lead scoring case study presentation
 
TAUS 2.0 and the Game Changers in Localization, by Jaap van der Meer, directo...
TAUS 2.0 and the Game Changers in Localization, by Jaap van der Meer, directo...TAUS 2.0 and the Game Changers in Localization, by Jaap van der Meer, directo...
TAUS 2.0 and the Game Changers in Localization, by Jaap van der Meer, directo...
 
Devops1
Devops1Devops1
Devops1
 
Agile transformation by Gnanasambandham anbazhagan
Agile transformation   by Gnanasambandham anbazhaganAgile transformation   by Gnanasambandham anbazhagan
Agile transformation by Gnanasambandham anbazhagan
 
Translation and localization process optimization - www.konsul.info
Translation and localization process optimization - www.konsul.infoTranslation and localization process optimization - www.konsul.info
Translation and localization process optimization - www.konsul.info
 
Building Large Arabic Multi-Domain Resources for Sentiment Analysis
Building Large Arabic Multi-Domain Resources for Sentiment Analysis Building Large Arabic Multi-Domain Resources for Sentiment Analysis
Building Large Arabic Multi-Domain Resources for Sentiment Analysis
 
Rapid Strategic SRE Assessments
Rapid Strategic SRE AssessmentsRapid Strategic SRE Assessments
Rapid Strategic SRE Assessments
 
Ncerc rlmca202 adm m4 ssm
Ncerc rlmca202 adm m4 ssmNcerc rlmca202 adm m4 ssm
Ncerc rlmca202 adm m4 ssm
 
Classification of webpages as Ephemeral or Evergreen
Classification of webpages as Ephemeral or EvergreenClassification of webpages as Ephemeral or Evergreen
Classification of webpages as Ephemeral or Evergreen
 
North Star Case Study.pptx
North Star Case Study.pptxNorth Star Case Study.pptx
North Star Case Study.pptx
 
TCS Interview Questions and Answers 2022 | How to Crack TCS Interview for Fre...
TCS Interview Questions and Answers 2022 | How to Crack TCS Interview for Fre...TCS Interview Questions and Answers 2022 | How to Crack TCS Interview for Fre...
TCS Interview Questions and Answers 2022 | How to Crack TCS Interview for Fre...
 
Applying a Quality Framework to the In-depth Intrerview Method
Applying a Quality Framework to the In-depth Intrerview MethodApplying a Quality Framework to the In-depth Intrerview Method
Applying a Quality Framework to the In-depth Intrerview Method
 
DevOps
DevOpsDevOps
DevOps
 
The Quest for Quality at Speed
The Quest for Quality at SpeedThe Quest for Quality at Speed
The Quest for Quality at Speed
 
Part of speech tagging for Arabic
Part of speech tagging for ArabicPart of speech tagging for Arabic
Part of speech tagging for Arabic
 

Recently uploaded

Analog to Digital and Digital to Analog Converter
Analog to Digital and Digital to Analog ConverterAnalog to Digital and Digital to Analog Converter
Analog to Digital and Digital to Analog ConverterAbhinavSharma374939
 
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...ranjana rawat
 
Introduction to Multiple Access Protocol.pptx
Introduction to Multiple Access Protocol.pptxIntroduction to Multiple Access Protocol.pptx
Introduction to Multiple Access Protocol.pptxupamatechverse
 
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINEMANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINESIVASHANKAR N
 
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLSMANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLSSIVASHANKAR N
 
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Serviceranjana rawat
 
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Dr.Costas Sachpazis
 
247267395-1-Symmetric-and-distributed-shared-memory-architectures-ppt (1).ppt
247267395-1-Symmetric-and-distributed-shared-memory-architectures-ppt (1).ppt247267395-1-Symmetric-and-distributed-shared-memory-architectures-ppt (1).ppt
247267395-1-Symmetric-and-distributed-shared-memory-architectures-ppt (1).pptssuser5c9d4b1
 
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxWhat are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxwendy cai
 
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur High Profile
 
Internship report on mechanical engineering
Internship report on mechanical engineeringInternship report on mechanical engineering
Internship report on mechanical engineeringmalavadedarshan25
 
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICSAPPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICSKurinjimalarL3
 
High Profile Call Girls Nashik Megha 7001305949 Independent Escort Service Na...
High Profile Call Girls Nashik Megha 7001305949 Independent Escort Service Na...High Profile Call Girls Nashik Megha 7001305949 Independent Escort Service Na...
High Profile Call Girls Nashik Megha 7001305949 Independent Escort Service Na...Call Girls in Nagpur High Profile
 
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)Suman Mia
 
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptxDecoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptxJoão Esperancinha
 
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...ranjana rawat
 
Processing & Properties of Floor and Wall Tiles.pptx
Processing & Properties of Floor and Wall Tiles.pptxProcessing & Properties of Floor and Wall Tiles.pptx
Processing & Properties of Floor and Wall Tiles.pptxpranjaldaimarysona
 

Recently uploaded (20)

Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptxExploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
Exploring_Network_Security_with_JA3_by_Rakesh Seal.pptx
 
Analog to Digital and Digital to Analog Converter
Analog to Digital and Digital to Analog ConverterAnalog to Digital and Digital to Analog Converter
Analog to Digital and Digital to Analog Converter
 
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANJALI) Dange Chowk Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
 
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
★ CALL US 9953330565 ( HOT Young Call Girls In Badarpur delhi NCR
 
Introduction to Multiple Access Protocol.pptx
Introduction to Multiple Access Protocol.pptxIntroduction to Multiple Access Protocol.pptx
Introduction to Multiple Access Protocol.pptx
 
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINEMANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
MANUFACTURING PROCESS-II UNIT-2 LATHE MACHINE
 
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
9953056974 Call Girls In South Ex, Escorts (Delhi) NCR.pdf
 
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLSMANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
MANUFACTURING PROCESS-II UNIT-5 NC MACHINE TOOLS
 
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
(RIA) Call Girls Bhosari ( 7001035870 ) HI-Fi Pune Escorts Service
 
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
Sheet Pile Wall Design and Construction: A Practical Guide for Civil Engineer...
 
247267395-1-Symmetric-and-distributed-shared-memory-architectures-ppt (1).ppt
247267395-1-Symmetric-and-distributed-shared-memory-architectures-ppt (1).ppt247267395-1-Symmetric-and-distributed-shared-memory-architectures-ppt (1).ppt
247267395-1-Symmetric-and-distributed-shared-memory-architectures-ppt (1).ppt
 
What are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptxWhat are the advantages and disadvantages of membrane structures.pptx
What are the advantages and disadvantages of membrane structures.pptx
 
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur EscortsCall Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
Call Girls in Nagpur Suman Call 7001035870 Meet With Nagpur Escorts
 
Internship report on mechanical engineering
Internship report on mechanical engineeringInternship report on mechanical engineering
Internship report on mechanical engineering
 
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICSAPPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
APPLICATIONS-AC/DC DRIVES-OPERATING CHARACTERISTICS
 
High Profile Call Girls Nashik Megha 7001305949 Independent Escort Service Na...
High Profile Call Girls Nashik Megha 7001305949 Independent Escort Service Na...High Profile Call Girls Nashik Megha 7001305949 Independent Escort Service Na...
High Profile Call Girls Nashik Megha 7001305949 Independent Escort Service Na...
 
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)Software Development Life Cycle By  Team Orange (Dept. of Pharmacy)
Software Development Life Cycle By Team Orange (Dept. of Pharmacy)
 
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptxDecoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
Decoding Kotlin - Your guide to solving the mysterious in Kotlin.pptx
 
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
(ANVI) Koregaon Park Call Girls Just Call 7001035870 [ Cash on Delivery ] Pun...
 
Processing & Properties of Floor and Wall Tiles.pptx
Processing & Properties of Floor and Wall Tiles.pptxProcessing & Properties of Floor and Wall Tiles.pptx
Processing & Properties of Floor and Wall Tiles.pptx
 

ISNCC '23 Presentation.pptx

  • 1. ARABIC SENTIMENT ANALYSIS O F F O O D DEL IVERY SERVICES REVIEWS DR. DHEYA MUSTAFA FACULTY OF ENGINEERING THE HASHEMITE UNIVERSITY Oct 24, 2023 ISNCC’2023, Doha, Qatar 1
  • 3. INTRODUCTION • Food delivery services (FDSs) have introduced diversity to the increasing demand for online food delivery marketplaces. • Mostly third-party marketplace apps for global ordering and delivery, such as UberEATS, Talabat (the Arabic word for orders), and Menulog. • Employ a cost intensive, aggregator business model and are in charge of all delivery logistics. 3
  • 4. MOTIVATION • The majority of businesses want to effectively increase customer satisfaction by using data to identify areas for improvement. –Mostly based on reviews from customers. • Sentiment analysis (SA) can determine the customers’ opinion based on the written content. 4
  • 5. GOAL • Apply SA to Arabic content to give an even better service, driven by the feedback of customers. • Offer a quick and efficient monitoring system to handle the needs clients on a large scale at a low cost and maximum profit. 5
  • 6. METHODOLOGY • gathering Arabic reviews regarding FDS and evaluating feelings using popular ML methods. • Created an Arabic dialects lexicon including various dialects and their standard orientation in Arabic. • Emotions are identified and analyzed using an existing emotion lexicon. • experiments were carried out before and after applying cleaning and preprocessing . • Negative sentiments were analyzed to identify the possible causes. 6
  • 7. 7
  • 8. DATASET CLEANING • The raw datasets contain many empty and repetitive reviews that should be efficiently eliminated. • Involves removing duplications, identifying emojis, characters, and word replacements, spelling correction, and annotation. • SA of dialectical Arabic is challenging. • We constructed a dialects dictionary manually to translate dialectical phrases into equivalent MSA in the context of FDS. • Popular English words written in the Arabic alphabet in the reviews are included in the dialects dictionary as well. – Example, the word “nice” is written as ‘‫نايس‬ ’ and is equivalent to the word ‘‫جميل‬ ’ in Arabic. 8
  • 9. 9
  • 10. ARABIC REVIEWS PREPROCESSING • Clean review texts are process using common NLP tasks. • normalization : represent step characters that have more than one form like ( ،‫أ‬ ،‫إ‬ ‫آ‬ ) in a unified form. • Also removed any punctuations, diacritics, and tatweel. • Stopword removal: reduce the size of the text by removing unnecessary particles and pronouns. • Tokenization: split the text into single units such as sentences and then into single words. 10
  • 11. ARABIC REVIEWS PREPROCESSING • Part of Speech (PoS) tagging: annotate words in a text based on their type and their relationships to neighboring and related words in that text. – Based on the Stanford tagger [35]. • The last NLP task involved removing suffixes or prefixes. – To efficiently convert each word in its base form without extensions (Stemming) – Used the tashaphyne stemmer 11
  • 12.
  • 13. SENTIMENT CLASSIFICATION • The final phase was to build and train the machine learning (ML) classifier using a labeled subset of the dataset (Talabat FDS ). • We tested the model on unseen reviews and categorized them into positive, negative, or neutral classes. • ML models: DT and SVM – Most frequently used classifiers in Arabic SA – Often provide the best results over other ML classifiers
  • 15. EVALUATION 15 Experiment DT classifier SVM Classifier P% R% F% Acc% P% R% F% Acc% Before cleaning and pre- processin g +ve 67 95 79 87 86 86 -ve 82 44 57 80 86 83 Neu 20 4 6 31 23 26 Avg 68 69 65 69 79 80 80 80 After cleaning and pre- processin g +ve 72 91 81 87 88 88 -ve 76 60 67 77 89 83 Neu 20 3 5 33 10 15 Avg 69 73 69 73 79 82 80 82 +ve: positive -ve: negative Neu: neutral Avg: average
  • 17. EVALUATION 17 Fig. 3. A confusion matrix of the SVM classifier after applying cleaning and preprocessing phases.
  • 19. CONCLUSION • Proposed a sentiment analysis model of Arabic dialect reviews about FDS. • applied intensive preprocessing and several aspects of analyzing Arabic text to generate a dataset. • examined two well-known ML classifiers over this dataset and provided a comparative performance evaluation • Analyzed the most representative negative reviews and produced a word cloud. – The long waiting time, cold food, wrong order, and bill were the most important factors contributing to the negative sentiments. 19
  • 20. Thank you Oct. 23, 2023 ISNCC'2014, Doha, Qatar 20

Editor's Notes

  1. Can be extended to cover more dialects and identify their standard orientations in Arabic or be customized to a specific domain.
  2. Each word was assigned a corresponding tag that indicates its role in the sentence.