SCIENCE@home

Hybrid African Buffalo Optimization with Convolutional Neural Network for Unstructured Social Media Data Processing and Classification using NLP

st H. Ramprasanth,M. Rahul

Published 2025 in 2025 5th International Conference on Mobile Networks and Wireless Communications (ICMNWC)

ABSTRACT

Real-time scenarios experience a continuous increase in data volume over time. The amount of data on social networking sites, such as Facebook and Twitter, has grew uncontrollably. The recommended approach would simultaneously process the enormous amount of data in distributed clusters as discrete data, aggregating all the data from each cluster to generate the data. With billions of users worldwide, social media has a significant influence on the big data space. The growing popularity of social media has resulted in the production of a substantial quantity of unstructured data through user-generated content, which includes text, audio, video, and web pages. The need of customers to know who is who and what is what in their daily lives cannot prevent the benefits of social network data analysis. Their classification was the outcome of social media’s unstructured data. Several machine learning classifiers that employed performance metrics such as accuracy, precision, sensitivity, and specificity were also evaluated. The effect of feature extraction methods on classification accuracy was also looked at. To examine the chosen dataset from the Kaggle website, Hybrid African Buffalo Optimization with Convolutional Neural Network (HABOCNN) was also employed. The HABOCNN outperforms the other classifiers, according to the results. Using TF and TF-IDF features, the HABOCNN can extract features with 94.14% and 94.14% accuracy, respectively. The outcomes show that ensemble classifiers are more accurate than non-ensemble classifiers. Further experiments demonstrate the superiority of TF-IDF as a feature extraction method for machine learning classifiers. TF and TF-IDF feature extraction perform better than word2vec feature extraction. The accuracy of the HACOANN is lower than that of ML classifiers.

PUBLICATION RECORD

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-20 of 20 references · Page 1 of 1

CITED BY

  • No citing papers are available for this paper.

Showing 0-0 of 0 citing papers · Page 1 of 1