Page 72 - IJOCTA-15-1

O. Ayana, D. F. Kanbak, M. Kaya Keles / IJOCTA, Vol.15, No.1, pp.50-70 (2025)

was utilized for feature selection, it primarily fo- Author contributions
cuses on word frequency and may overlook seman-
tic relationships and contextual nuances. This Conceptualization: Omer Ayana, Deniz Furkan
limitation could potentially impact the accuracy Kanbak
of the sentiment classification. Furthermore, the
Formal analysis: All authors
use of a standard optimization algorithm, rather
Investigation: All authors
than a hybrid variant, may limit the exploration
Methodology: Omer Ayana, Deniz Furkan Kan-
of more sophisticated or effective methodologies
bak
that could yield better performance. Finally,
Writing – original draft: Omer Ayana, Deniz
the study’s sentiment classification is constrained
Furkan Kanbak
to binary categories, positive and negative, po-
Writing – review & editing: Mumine Kaya Keles
tentially missing finer distinctions among various
emotional states.
Availability of data
Future research could address these limitations
by incorporating a broader range of data cat- Not applicable.
egories, thereby improving the generalizability
of the findings. Employing hybrid optimization References
techniques may also enhance the performance of
the SA by leveraging the strengths of multiple al- [1] Choudhary, M., & Choudhary, P.K. (2018). Sen-
gorithms. Moreover, exploring advanced method- timent analysis of text reviewing algorithm us-
ologies, such as semantic analysis approaches, ing data mining. 2018 International Conference
could provide a more nuanced understanding of on Smart Systems and Inventive Technology (IC-
sentiment expressions. Furthermore, expanding SSIT), 532–538.
the scope of sentiment classification to include [2] Khuc, V.N., Shivade, C., Ramnath, R., & Ra-
more specific emotional categories, such as joy, manathan, J. (2012). Towards building large-scale
anger, or sadness, would allow for a deeper analy- distributed systems for twitter sentiment analysis.
Proceedings of the 27th annual ACM symposium
sis of user feedback and contribute to more effec-
on applied computing, 459–464.
tive product development and customer satisfac-
[3] Subramaniyaswamy, V., Vijayakumar, V., Lo-
tion strategies. Finally, the BSO algorithm can
gesh, R., & Indragandhi, V. (2015). Unstructured
be effectively utilized for selecting optimal key-
data analysis on big data using map reduce. Pro-
words in website and article classification prob- cedia Computer Science, 50, 456–465.
lems. It also serves to optimize the inter-word [4] Kumar, K.L. Santhosh, Desai, J., & Majumdar,
weights for word embeddings. Additionally, BSO J. (2016). Opinion mining and sentiment analysis
can be applied to diagnose conditions based on re- on online customer review. 2016 IEEE Interna-
ports generated from medical imaging techniques tional Conference on Computational Intelligence
such as ultrasound, PET, and MRI. and Computing Research (ICCIC), 1–4.
[5] Tang, X., Hao, B., Dang, X., Zhong, B., Wang,
R., & Yan, Z. (2020). Text semantic understand-
ing based on knowledge enhancement and multi-
Acknowledgments granular feature extraction. 2020 Chinese Au-
tomation Congress (CAC), 337–341.
[6] Guzman, E., & Maalej, W. (2014). How do
None. users like this feature? A fine-grained sentiment
analysis of app reviews. 2014 IEEE 22nd In-
ternational Requirements Engineering Conference
(RE), 153–162.
Funding
[7] Kumar, A., Srinivasan, K., Cheng, W.-H., &
Zomaya, A.Y. (2020). Hybrid context enriched
None. deep learning model for fine-grained sentiment
analysis in textual and visual semiotic modality
social data. Information Processing & Manage-
ment, 57(1), 102141.
Conflict of interest [8] Shi, H., Zhou, G., & Qian, P. (2010). An
attribute-based sentiment analysis system. Infor-
mation Technology Journal, 9(8), 1607–1614.
The authors declare that they have no conflict of [9] Zeng, J., Li, F., Liu, H., Wen, J., & Hirokawa, S.
interest regarding the publication of this article. (2016). A restaurant recommender system based
66