Тилини ривожлантириш департаменти ўзбекистон республикаси олий ва ўрта махсус таълим вазирлиги

davlat o‘zbek tili va adabiyoti 
universiteti 
“O‘ZBEK MILLIY VA TA’LIMIY 
KORPUSLARINI YARATISHNING NAZARIY 
HAMDA AMALIY MASALALARI”
Xalqaro ilmiy-amaliy konferensiya
Vol. 1
№. 01 (2021) 
292 
Different well-known machine learning algorithms are used in the classification stage. In particular, 
we used Instance Based Classifier (IBk), Neural Networks (NN), Support Vector Machines (SVM-rbf, 
SVM-poly), Bayesian Classifier (BN), Decision Trees algorithms C4.5 (J48), Random Forest (RF), Fast 
Decision Tree Learner (RT). All classification algorithms were implemented using WEKA software. For 
all algorithms, the free parameters were empirically selected, while parameter values not reported here 
were kept in their default values. For training and testing classification algorithms, 10-fold cross-
validation was applied to the UMR dataset. The results obtained are presented in the table below. 
Table 1. Opinion classification accuracy for different classification algorithms 
Classification 
algorithms 
IBk 
NN 
SVM-
poly 
SVM-
rbf 
J48 
RF 
RT 
BN 
Accuracy (%) 
80.26 
82.72 
84.55 
84.39 
83.46 
85.25 
84.12 
75.34 
As can be seen in this table, the best classification accuracy was 85.25% for the random forest 
algorithm, while the SVM-poly, SVM-rbf, and RT algorithms achieved around 1% lower accuracy than 
the random forest algorithm. 
REFERENCES: 
1.
El-Masri M., Altrabsheh N. and Mansour H. Successes and challenges of Arabic sentiment analysis 
research: a literature review. Social Network Analysis and Mining, 7(1), 2017. – pp.1-22.
2.
Mikolov T., Sutskever I., Chen K., Corrado G.S., Dean J. Distributed representations of words and 
phrases and their compositionality. Advances in neural information processing systems, 2013. – 
P.3111-3119. 
3.
Pennington J., Socher R. and Manning C. Glove: global vectors for word representation. Proceedings 
of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), 2014. – P. 
1532-1543. 
4.
Bojanowski P., Grave E., Joulin A. and Mikolov T. Enriching word vectors with subword 
information. Transactions of the Association for Computational Linguistics, 5, 2017. – P.135-146. 
5.
Rabbimov I., Mporas I., Simaki V., Kobilov S. Investigating the Effect of Emoji in Opinion 
Classification of Uzbek Movie Review Comments. Volume 12335 of the series Lecture Notes in 
Artificial Intelligence. Springer, 2020. 
6.
Kuriyozov E., Matlatipov S. Building a New Sentiment Analysis Dataset for Uzbek Language and 
Creating Baseline Models. 
Proceedings
. 2019; 21(1). – p.37.
7.
Rabbimov I., Kobilov S. Multi-Class Text Classification of Uzbek News Articles using Machine 
Learning. In Journal of Physics: Conference Series, 2020, Vol. 1546. – P. 012097.
8.
Rabbimov I., Kobilov S. and Mporas I. Uzbek News Categorization using Word Embeddings and 
Convolutional Neural Networks. 2020 IEEE 14th International Conference on Application of 
Information and Communication Technologies (AICT), Tashkent, Uzbekistan, 2020. – P.1-5. 

---