Authors

  • Malika Suyunova
    Tashkent State University named after Alisher Navoi

DOI:

https://doi.org/10.71337/inlibrary.uz.jasss.76698

Abstract

This article examines the issues of sentiment analysis, which is a branch of computational linguistics, and its study in the context of world linguistics. Computational linguistics is a scientific field focused on understanding and processing human language with the help of computers, and the importance of sentiment analysis is growing day by day. By detecting and classifying emotional states in text, it is possible to analyze various practical areas, including social networks, customer reviews, and other texts. The article provides detailed information on the main methods and techniques of sentiment analysis, as well as approaches to its study in linguistics. It emphasizes the importance of the semantic characteristics of language, contextual analysis, and cross-lingual research.

 

 

background image

Volume 15 Issue 03, March 2025

Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:

6.995, 2024 7.75

http://www.internationaljournal.co.in/index.php/jasass

891

SENTIMENT ANALYSIS RESEARCH AND IMPORTANCE

Suyunova Malika Adil kizi

Tashkent State University named after Alisher Navoi

University of Uzbek Language and Literature

Computational linguistics and digital technologies

PhD student of the department

Email:

malikasuyunova0@gamil.com

Annotation:

This article examines the issues of sentiment analysis, which is a branch of

computational linguistics, and its study in the context of world linguistics. Computational

linguistics is a scientific field focused on understanding and processing human language with the

help of computers, and the importance of sentiment analysis is growing day by day. By detecting

and classifying emotional states in text, it is possible to analyze various practical areas, including

social networks, customer reviews, and other texts. The article provides detailed information on

the main methods and techniques of sentiment analysis, as well as approaches to its study in

linguistics. It emphasizes the importance of the semantic characteristics of language, contextual

analysis, and cross-lingual research.

Keywords:

sentiment analysis, language processing, emotional analysis, natural language

processing (NLP), cross-lingual research, semantics, linguistics, social networks, artificial

intelligence, grammatical analysis.

Аннотация:

Эта статья изучает вопросы компьютерной лингвистики, связанные с

анализом сентимента и его исследованием в контексте мировой лингвистики.

Компьютерная лингвистика — это научная область, направленная на понимание и

обработку человеческого языка с помощью компьютеров, и значение анализа сентимента

с каждым днем растет. Определение эмоционального состояния в тексте и его

классификация позволяют проводить анализ в различных практических областях, включая

социальные сети, отзывы клиентов и другие тексты. В статье представлены основные

методы и подходы анализа сентимента, а также подходы к его исследованию в

лингвистике. В частности, подчеркивается важность семантических характеристик языка,

контекстуального анализа и кросс-языковых исследований.

Ключевые слова:

анализ сентимента, обработка языка, эмоциональный анализ,

естественная обработка языка (NLP), кросс-языковые исследования, семантика,

лингвистика, социальные сети, искусственный интеллект, грамматический анализ.

Introduction.

Today, the demand for information technologies is increasing day by day. All

industries, enterprises and companies in the world have been covered by computer technologies.

As technologies develop, competition is also increasing by promoting products and services on

social networks. The increase or decrease in demand for a product depends on whether the

opinions given to it on social networks are positive or negative. Today, the flow of feedback on

products and services on social networks is increasing immensely. This requires a lot of


background image

Volume 15 Issue 03, March 2025

Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:

6.995, 2024 7.75

http://www.internationaljournal.co.in/index.php/jasass

892

manpower to determine whether each expressed opinion is positive, negative or neutral. The

human factor cannot analyze this level of large information in one day or several days. However,

the number of data and comments increases every hour. Solving these and similar language-

related issues using computer models is one of the current issues in the field of natural language

processing (NLP). Analyzing the feelings expressed by a person to a particular product, service

or brand is called sentiment analysis. This is one of the most active research areas in natural

language processing. Sentiment analysis is one of the important areas in the field of detecting

emotional states in text and computational linguistics. This method serves to identify positive,

negative or neutral emotions expressed in texts and on this basis, analyzes texts in various fields,

including customer reviews, social networks, advertising materials and other texts. The study of

sentiment analysis is also widespread in world linguistics, and research in this area helps to

understand not only analysis methods, but also the semantic, syntactic and contextual features of

language [1].

The main goal of sentiment analysis is to determine the emotional state of a text or word. For

example, a customer review of a product indicates a positive or negative attitude. Sentiment

analysis uses computational linguistics technologies to analyze the position, grammatical

structure, and meaning of words in the text [2].

Sentiment analysis falls into three main categories:

1.

Positive: The essence of the text or word conveys a good or positive feeling.

2.

Negative: The text or word expresses negativity or disapproval.

3.

Neutrality: The emotional content of a text or word is either unclear or emains neutral.

Various methods help perform this analysis:

Word-based analysis: This method determines the emotional image (positive or negative)

of each word.

Phrase and context-based analysis: In a full text analysis, the context and grammatical

structure between words are considered.

Machine learning technologies: Using artificial intelligence and deep learning techniques,

the accuracy and efficiency of sentiment analysis are increased.

Linguists analyze the semantic (meaning) aspects of words in different languages. For sentiment

analysis, it is important to understand the specific features of this language, because each

language has its own grammatical and lexical features. Since the semantic and lexical systems of

each language are different, translinguistic (cross-linguistic) studies of sentiment analysis are of

great importance in world linguistics. Linguists study how to identify positive and negative

words in several languages, and how these words change in cultural context. Sentiment analysis,

especially if there are differences between languages, can give erroneous results if performed

solely based on words. Therefore, in world linguistics, focus is placed on analyzing the text as a

whole, because the meaning of words depends on their context, structure, and cultural features.

Sentiment analysis studies not only the structure of the language, but also the author's intention,

his cultural and sociological context [3]. Linguists are also studying how accurate and reliable

sentiment analysis can be made of sentiments that emerge on social media or online platforms.

The study of sentiment analysis in the field of computational linguistics forms an important

bridge between linguistics and computer science. This field provides the opportunity to analyze

the natural structure of language using computers, to identify emotional states between texts, and

to enhance international interactions. For world linguistics, sentiment analysis helps to


background image

Volume 15 Issue 03, March 2025

Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:

6.995, 2024 7.75

http://www.internationaljournal.co.in/index.php/jasass

893

understand the emotional and contextual factors of language, creating opportunities for cross-

linguistic research, cultural and social analysis.

The significance of the study.

This study aims to provide an in-depth analysis of scientific

research and practical applications in the fields of computational linguistics and sentiment

analysis. Today, with the development of technology and artificial intelligence, the capabilities

of sentiment analysis are expanding, and its analytical and scientific aspects are increasing as

well. The processes of analyzing texts, identifying emotional states and classifying them using

computational linguistics are widely used in various fields, including marketing, social networks,

customer opinion research, and obtaining feedback on products. The importance of the study

consists in showing how sentiment analysis has bin studied in world linguistics and the role of

this analysis in determining the semantic and contextual features of language. Deepening the

interaction between linguistics and computer science and developing cross-language research

will serve the advancement of modern linguistics and language processing technologies. At the

same time, how sentiment analysis differs across different languages and cultures and its impact

on scientific and applied fields increases the relevance of the research.
Today, sentiment analysis of Uzbek texts is one of the most relevant issues. This, in turn, allows

using this data for various purposes, including marketing, politics, social research, and other

areas. Sentiment analysis allows you to determine not only the general emotional direction of the

text, but also the author's attitude to a particular object or topic.

Literature review.

A number of important studies have emerged in the fields of computational

linguistics and sentiment analysis in recent years. These studies have studied various methods

and practical applications for automatic language analysis and emotional state detection. The

literature review reviews the main theoretical and practical aspects of computational linguistics

and sentiment analysis, and provides some important information about the development and

application possibilities of these fields. Computational linguistics mainly focuses on automatic

language analysis. The first work in this field began in the 1950s and gained momentum in the

following years by the rapid development of deep learning and artificial intelligence

technologies. Studies by Allen (1995), Jurafsky and Martin (2020) provide a lot of important

information about computational linguistics methods, including natural language processing

(NLP), and its practical applications. These methods serve to analyze texts, identify grammatical

structures, and extract semantic meanings [4, 5].

In the world, several studies have been conducted on building a language model and sentiment

analysis based on artificial intelligence algorithms, as well as developing software modules and

tools in the process of natural language processing. The theoretical foundations of sentiment

analysis are described in the works of S.Yu.Toldova, AA.Bonch-Osmolovskaya, T.Sadikov.

Among them, Stanford University professors Christopher Manning, Dan Jurafsky and Percy

Liang (Natural Language Processing Group, USA), University of Southern California professors

Robin Jia and Jesse Thomason (NLP Group, USA), University of Oregon professor Liang Huang

(NLP Group, USA) are conducting research. In addition, the work of Heidelberg University

(Germany) and others in this area has undergone comparative analysis. Stanford University

professor Christopher Manning's contribution to the field of natural language processing has


background image

Volume 15 Issue 03, March 2025

Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:

6.995, 2024 7.75

http://www.internationaljournal.co.in/index.php/jasass

894

shown the effectiveness of advanced linguistic models and machine learning methods, and has

yielded high results in SVM-based sentiment analysis. His research has significantly advanced

the understanding and application of sentiment analysis, emphasizing the importance of complex

algorithms and comprehensive linguistic concepts in correctly analyzing textual data [6]. The

development of sentiment analysis began in the 2000s and quickly gained its practical

application. Pang and Lee (2008) presented the main methods of sentiment analysis, including

word- and phrase-based analysis, in detail. The work of Ian Maas, Ray Yeung, and Richard PL

(2011) on the IMDB dataset is regarded one of the key works in the field of sentiment analysis.

They used deep learning methods such as RNN (recurrent neural networks) and SNN

(convolutional neural networks) [7]. Socher et al. (2013) created the Stanford Sentiment

Treebank and used this dataset for sentiment analysis. They further developed the analysis of

texts with a certain structure using RNN and its variations (e.g. LSTM) [8]. Zadeh et al. (2016)

created the SMU-MOSI dataset and studied multimodal sentiment analysis. They combined text,

image, and audio data and performed multi-factor analysis [9].

While early approaches to sentiment analysis focused on analyzing the emotional value of words,

later deep learning and machine learning technologies played a key role in understanding the full

context of the text (Devlin et al., 2018). Sentiment analysis is one of the most complex problems

in linguistics, as each language has its own semantic and lexical structures [10]. Therefore, cross-

language sentiment analysis research is of great importance. A study by Shen et al. (2018) aimed

to determine the effectiveness of cross-language sentiment analysis and the differences between

different languages. In this study, the issues of how the semantic features of one language affect

another language and how to implement translinguistic analysis in linguistics were investigated

[11].

The study of sentiment analysis in linguistics includes not only semantic and lexical analysis, but

also the social, cultural and pragmatic aspects of language. The research presented by Hoveyda

and Falk (2017) aimed to study the cultural and social context of sentiment analysis [12]. They

showed how sentiment analysis identifies emotional states in each culture and the impact of

different languages and cultures on sentiment analysis. Cross-language (interlingual) studies of

sentiment analysis have been undertaken. In this method, researchers examined the process of

sentiment identification in several languages. The lexical and semantic features of each language

and its cultural context were factored in. Cross-language sentiment analysis analyzed the

differences and similarities of positive and negative words that exist in different languages. A

series of experiments were conducted to measure the effectiveness of sentiment analysis. In these

experiments, models were built based on the above methods to identify different types of

sentiment (positive, negative, neutral) in texts. The results were tested and their accuracy and

efficiency were evaluated [13, 14]. During the experiments, the model was tested taking into

account different lengths, topics, and languages of the texts.

A number of scientific studies on sentiment analysis are also being conducted in the Uzbek

language. For example, S. Matlatipov worked on the topic “Sentiment analysis of the Uzbek

language”. He developed methods for intellectual analysis of the given data to solve the problem

of sentiment analysis of proposals and opinions written in digital text in the Uzbek language [15].

A group of researchers led by Ilyos Rabbimov at Samarkand State University studied the

comments left on Uzbek films based on emojis. S. Allanazarova, a graduate student at the


background image

Volume 15 Issue 03, March 2025

Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:

6.995, 2024 7.75

http://www.internationaljournal.co.in/index.php/jasass

895

National University of Uzbekistan named after Mirzo Ulugbek, discussed the issue of studying

texts from an emotional perspective.

Discussion.

This study investigated computational linguistics and sentiment analysis methods.

The main objective of the study was to improve the efficiency of sentiment analysis and to

identify the most suitable approach by comparing different methods. During the study, various

methods such as word-based analysis, machine learning, deep learning, and cross-language

analysis were tested. The advantages and disadvantages of each method were analyzed, and ways

to improve the efficiency of sentiment analysis were identified. The simplicity and speed of this

method are one of its greatest advantages. However, its disadvantage is that it does not take into

account contextual information in the text. This can sometimes lead to errors, since one word can

have different meanings and emotions in different contexts. Therefore, word-based methods

alone are not enough to improve the accuracy and efficiency of sentiment analysis [16].

Deep learning techniques, such as advanced networks such as RNN, LSTM or Transformer,

provide accurate and efficient implementation of sentiment analysis. They allow for a complete

understanding of the contextual meanings and semantic structure of the text. However, these

techniques require large computational resources and can sometimes be highly complex. Also

known as, training these techniques requires time and effort. Cross-language sentiment analysis

is very useful when working with different languages and can be widely used on a global scale.

These techniques can be used to analyze texts in multiple languages and identify emotional states

in different cultures. However, there are difficulties in taking into account the specific lexical

and semantic features of the language. The unique structure of each language, including

differences between words and phrases, complicates sentiment analysis.

Each method has its own advantages and disadvantages, and when compared with each other,

each works effectively in its own specific situations. Word-based analysis is good for simple

tasks, but deep learning and machine learning methods provide higher accuracy when analyzing

complex and long texts. Cross-language analysis is important when working in different

languages and cultural contexts. For future research, it is necessary to improve the accuracy of

sentiment analysis and develop new approaches to integrating different methods. For example,

combining word-based analysis and machine learning methods can better identify contextual and

lexical features of the text [17]. Also known as, through cross-language research, it is possible to

develop universal sentiment analysis systems in different languages and cultures. In general,

sentiment analysis is an important method that is widely used in the fields of computational

linguistics and linguistics today, and its scientific and practical importance is increasing day by

day. Therefore, it is necessary to conduct more in-depth research in this area and introduce

advanced technologies, which will not only improve the process of understanding language, but

also allow for its effective use in many social and economic areas.

Abstract.

This study aims to study various methods in the field of computational linguistics and

sentiment analysis and compare their effectiveness. The study extensively analyzed methods

such as word-based analysis, machine learning, deep learning, and cross-language analysis. The

advantages and disadvantages of each method were identified and their effectiveness in different

situations was shown. Word-based analysis methods, although they work quickly and easily, do

not take into account the context of the text, and therefore can sometimes lead to errors. Machine

learning methods provide accurate results based on more data, but they require a large amount of

training data for their effective operation. Deep learning methods, in particular Transformer


background image

Volume 15 Issue 03, March 2025

Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:

6.995, 2024 7.75

http://www.internationaljournal.co.in/index.php/jasass

896

networks, are very effective in analyzing the exact and complete context of texts, providing high

accuracy, but these methods require a lot of computational resources. Cross-language sentiment

analysis is effective in working with multiple languages and allows us to take into account

different cultures, but it may face some difficulties in taking into account differences between

languages. The results of the study show that the effectiveness of sentiment analysis depends on

the chosen method and the characteristics of the text being analyzed. Therefore, using a

combination of different methods helps to increase the accuracy of sentiment analysis. In future

research, it is possible to create more effective sentiment analysis systems by integrating

methods and developing new approaches, in particular, by jointly applying cross-language

analysis and deep learning technologies.

List of used literature

1.

Ismoilov, R. (2020). Computational Linguistics: Theoretical Foundations and

Practical Applications. Academy of Sciences of the Republic of Uzbekistan.

2.

Sattorov, F. (2018). Methodology of text processing and sentiment analysis.

Linguistics and Computational Linguistics, 12(3), 45-59.

3.

Azimov, M., & Akhmedov, S. (2021). Linguistics and sentiment analysis: New

approaches and research. In Uzbek Linguistics Journal 15(2), 118-134.

4.

Allen, J. (1995). Natural Language Understanding. Addison-Wesley.

5.

Jurafsky, D., & Martin, JH (2020). Speech and Language Processing: An Introduction

to Natural Language Processing, Computational Linguistics, and Speech Recognition. Pearson.

6.

G.S.Matlatipov. SENTIMENTAL ANALYSIS

OF THE UZBEKISTAN

LANGUAGE. Abstract. 6 p.

7.

Andrew L. Maas, Raymond E. Daly, Peter T. Pham, Dan Huang, Andrew Y. Ng, and

Christopher Potts. (2011).

Learning Word Vectors for Sentiment Analysis.

The 49th Annual

Meeting of the Association for Computational Linguistics (ACL 2011).b

8.

Richard Socher, Alex Perelygin, Jean Y. Wu, Jason Chuang, Christopher D. Manning,

Andrew Y. Ng and Christopher Potts (2013).

Recursive Deep Models for Semantic

Compositionality Over a Sentiment Treebank.

9.

Zadeh, A., Chen, M., Liang, J., Poria, S., & Morency, LP (2016). CMU-MOSI: A

Multimodal Sentiment Analysis Dataset.

Proceedings of the 2016 IEEE International

Conference on Data Mining (ICDM)

, 2016, 1-10. https://doi.org/10.1109/ICDM.2016.0042.

10.

Devlin, J., Chang, M., Lee, K., & Toutanova, K. (2018). BERT: Pre-training of Deep

Bidirectional Transformers for Language Understanding. NAACL-HLT.

11.

Shen, D., Hu, Y., & Sun, Y. (2018). Cross-lingual Sentiment Analysis: A Survey.


background image

Volume 15 Issue 03, March 2025

Impact factor: 2019: 4.679 2020: 5.015 2021: 5.436, 2022: 5.242, 2023:

6.995, 2024 7.75

http://www.internationaljournal.co.in/index.php/jasass

897

IEEE Access, 6, 10506-10516.

12.

Hovy, D., & Falk, C. (2017). Sentiment Analysis and Its Application in Language and

Culture. Journal of Linguistic Research.

13.

Pang, B., & Lee, L. (2008). Opinion Mining and Sentiment Analysis. Foundations

and Trends in Information Retrieval, 2(1-2), 1-135.

14.

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A., Kaiser, Ł.,

& Polosukhin, I. (2017). Attention is All You Need. NIPS.

15.

G.S.Matlatipov. SENTIMENTAL ANALYSIS

OF THE UZBEKISTAN

LANGUAGE. Abstract.

16.

Davronov, A. (2020). Application of Computational Linguistics and Sentiment

Analysis in the Uzbek Language. Uzbekistan Institute of Science and Technology.

17.

Karimov, M., & Tashpulatov, D. (2021). Computational Linguistics and Its Use in

Different Languages: Practical Applications of Sentiment Analysis. International Scientific

Developments 3(5), 44-53.

References

Ismoilov, R. (2020). Computational Linguistics: Theoretical Foundations and Practical Applications. Academy of Sciences of the Republic of Uzbekistan.

Sattorov, F. (2018). Methodology of text processing and sentiment analysis. Linguistics and Computational Linguistics, 12(3), 45-59.

Azimov, M., & Akhmedov, S. (2021). Linguistics and sentiment analysis: New approaches and research. In Uzbek Linguistics Journal 15(2), 118-134.

Allen, J. (1995). Natural Language Understanding. Addison-Wesley.

Jurafsky, D., & Martin, JH (2020). Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition. Pearson.

G.S.Matlatipov. SENTIMENTAL ANALYSIS OF THE UZBEKISTAN LANGUAGE. Abstract. 6 p.

Andrew L. Maas, Raymond E. Daly, Peter T. Pham, Dan Huang, Andrew Y. Ng, and Christopher Potts. (2011). Learning Word Vectors for Sentiment Analysis. The 49th Annual Meeting of the Association for Computational Linguistics (ACL 2011).b

Richard Socher, Alex Perelygin, Jean Y. Wu, Jason Chuang, Christopher D. Manning, Andrew Y. Ng and Christopher Potts (2013). Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank.

Zadeh, A., Chen, M., Liang, J., Poria, S., & Morency, LP (2016). CMU-MOSI: A Multimodal Sentiment Analysis Dataset. Proceedings of the 2016 IEEE International Conference on Data Mining (ICDM) , 2016, 1-10. https://doi.org/10.1109/ICDM.2016.0042.

Devlin, J., Chang, M., Lee, K., & Toutanova, K. (2018). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. NAACL-HLT.