Desktop version

Home arrow Computer Science

  • Increase font
  • Decrease font


<<   CONTENTS   >>

RESULT AND DISCUSSION_

The evaluation of the proposed calculation focuses on a trial study of the most significant realities. To begin with, OGFIOCNN offered preferable results over conventional detection of cyberbullying. In addition, a content-based position was updated to evaluate various indicators such as failure and review for emotional evaluation. All tests were conducted using a Windows PC with 12 GB RAM and the execution was completed using MATLAB 2012.

Evaluation Metrics

Since recognition of cyberbullying is an ordering task, position accuracy is considered as the undisputed measurement solution. In any case, this is an unequal class problem. As a result, if accuracy is only considered as a measure, it is possible to achieve 80% accuracy, essentially identifying all test tweets with a large class of parts. This problem was understood by considering two different dimensions: visibility and precision. All measurements were recorded concurrently.

At this time, a complete correlation was made between the next three discoveries about cyberbullying. The goal is to show that the proposed calculation of OGHOCNN improves the current state of recognition of cyberbullying and it offers better expectations (greater accuracy), despite eliminating the need to create opportunities. The progress of the analysis begins with testing various qualities of channels, centers, pools and neurons to demonstrate that a change in the quality changes the nature of prediction. Table 10.1 shows the performance analysis of statistical measurements such as accuracy, precision, and recall in the proposed approach.

Figure 10.3 and Table 10.1 shows that the output of the proposed method, it is visible that the accuracy was 97%, precision was 98%, and recall was 97%. Figure 10.4 shows the comparison of proposed approach with existing accuracy measures, Figure 10.5 shows precision measure and Figure 10.6 shows recall measure of the proposed system.

TABLE 10.1 Performance Measures of the Proposed Research

No. of Neurons

Accuracy (%)

Precision (%)

Recall (%)

4

70

98

80

16

97

97

92

32

97

93

83

43

97

87

97

54

96

90

90

Graphical representation of proposed evaluation measures

FIGURE 10.3 Graphical representation of proposed evaluation measures.

Graphical representation of the proposed and existing accuracy measures

FIGURE 10.4 Graphical representation of the proposed and existing accuracy measures.

Graphical representation of proposed and existing precision measures

FIGURE 10.5 Graphical representation of proposed and existing precision measures.

TABLE 10.2 Comparison of the Proposed and Existing Accuracy Measures

No. of Neurons

OGHOCNN Accuracy (%)

SVM

NN

4

80

70

75

16

97

85

88

32

97

89

90

43

97

91

89

54

96

85

87

TABLE 10.3 Comparison of the Proposed and Existing Precision Measures

No. of Neurons

OGHOCNN Precision (%)

SVM

NN

4

98

90

91

16

97

89

88

32

93

85

86

43

87

79

78

54

90

85

87

TABLE 10.4 Comparison of Proposed and Existing Recall Measures

No. of Neurons

OGHOCNN Recall (%)

SVM

NN

4

80

75

77

16

92

85

87

32

83

77

76

43

97

89

85

54

90

85

88

Graphical representation of proposed and existing recall measures

FIGURE 10.6 Graphical representation of proposed and existing recall measures.

Comparative Analysis

The OGHOCNN method suggests that the characteristic work of Twitter dataset is useful in comparing SVM and NN techniques. The proposed representation of Twitter tweets dataset was analyzed by modifying the classifier, and the recorded results are as follows:

The time taken by various classifiers to predict the detection of cyberbullying is shown in Tables 10.2-10.4. The experimental results are presented in connection with the proposed prediction for cyberbullying detection using comparative classification classifier, OGHOCNN. Machine Learning Ability and Selectivity OGHOCNN describe the best accuracy, precision and recall activity. The precise accuracy measures determine the appropriateness of cyberbullying detection technology. The refresh rate was accurate, while the prediction for classification of a cyberbullying detection-based ranking model (RM) classifier was highly reliable. The accuracy, precision, and recall percentages for FP and FN coefficients got improved.

CONCLUSION

This research work proposed the design and development of classification RM based on supervised feature selection to detect cyberbullying on Twitter. This is aimed at improving the accuracy of detection and reduction in the execution times. The selection of supervised functions was carried out using the classification strategy. OGHOCNN was used to perform detection. The dataset was collected from Twitter tweets. The OGHOCNN classifier was implemented in MATLAB while the results offered better precision.

REFERENCES

  • 1. R. Pawar and R. R. Raje, “Multilingual Cyberbullying Detection System,” 2019 IEEE International Conference on Electro Information Technology (EIT), Brookings, SD, 2019, pp. 40-44.
  • 2. S. Parime and V. Suri, “Cyberbullying Detection and Prevention: Data Mining and Psychological Perspective,” 2014 International Conference on Circuits, Power and Computing Technologies [ICCPCT-2014], Nagercoil, 2014, pp. 1541-1547.
  • 3. A. Kovacevic, “Cyberbullying Detection Using Web Content Mining,” 2014 22nd Telecommunications Forum Telfor (TELFOR 2014), Belgrade, 2014, pp. 939-942.
  • 4. H. Rosa, J. P. Carvalho, P. Calado, B. Martins, R. Ribeiro and L. Coheur, “Using Fuzzy Fingerprints for Cyberbullying Detection in Social Networks,” 2018 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), Rio de Janeiro, 2018, pp. 1-7.
  • 5. I. Ting, W. S. Liou, D. Liberona, S. Wang and G. M. Tarazona Bermudez, “Towards the Detection of Cyberbullying based on Social Network Mining Techniques,” 2017 International Conference on Behavioral, Economic, Socio-Cultural Computing (BESC2017), Krakow, 2017, pp. 1-2.
  • 6. S. Tomkins, L. Getoor, Y. Chen and Y. Zhang, “A Socio-Linguistic Model for Cyberbullying Detection,” 2018 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2018), Barcelona, 2018, pp. 53-60.
  • 7. X. Zhang et al., “Cyberbullying Detection with a Pronunciation Based Convolutional Neural Network,” 2016 15th IEEE International Conference on Machine Learning and Applications (ICMLA 2016), Anaheim, CA, 2016, pp. 740-745.
  • 8. R. Zhao and К. Mao, “Cyberbullying Detection Based on Semantic-Enhanced Marginalized Denoising Auto-Encoder,” IEEE Transactions on Affective Computing, vol. 8, no. 3, pp. 328-339, 1 July-Sept. 2017.
  • 9. A. Upadhyay, A. Chaudhari, Arunesh, S. Ghale and S. S. Pawar, “Detection and Prevention Measures for Cyberbullying and Online Grooming,” 2017 International Conference on Inventive Systems and Control (ICISC 2017), Coimbatore, 2017, pp. 1-4.
  • 10. M. Yao, C. Chelmis and D. Zois, “Cyberbullying Detection on Instagram with Optimal Online Feature Selection,” 2018 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2018), Barcelona, 2018, pp. 401-408.
  • 11. D. Zois, A. Kapodistria, M. Yao and C. Chelmis, “Optimal Online Cyberbullying Detection,”
  • 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2018), Calgary, AB, 2018, pp. 2017-2021.
  • 12. J. Zhang, T. Otomo, L. Li and S. Nakajima, “Cyberbullying Detection on Twitter Using Multiple Textual Features,” 2019 IEEE 10th International Conference on Awareness Science and Technology (iCAST2019), Morioka, Japan, 2019, pp. 1-6.
  • 13. R. Sugandhi, A. Pande, S. Chawla, A. Agrawal and H. Bhagat, “Methods for Detection of Cyberbullying: A Survey,” 2015 15th International Conference on Intelligent Systems Design and Applications (ISDA 2015), Marrakech, 2015, pp. 173-177.
  • 14. Y. J. Foong and M. Oussalah, “Cyberbullying System Detection and Analysis,” 2017European Intelligence and Security Informatics Conference (EISIC 2017), Athens, 2017, pp. 40-46.
  • 15. M. A. Al-Ajlan and M. Ykhlef, “Optimized Twitter Cyberbullying Detection Based on Deep Learning,” 2018 21st Saudi Computer Society National Computer Conference (NCC 2018), Riyadh, 2018, pp. 1-5.
  • 16. V. Banerjee, J. Telavane, P. Gaikwad and P. Vartak, “Detection of Cyberbullying Using Deep Neural Network,” 2019 5th International Conference on Advanced Computing & Communication Systems (ICACCS 2019), Coimbatore, India, 2019, pp. 604-607.
  • 17. T. Mahlangu, C. Tu and P. Owolawi, “A Review of Automated Detection Methods for Cyberbullying,” 2018 International Conference on Intelligent and Innovative Computing Applications (ICONIC2018), PlaineMagnien, 2018, pp. 1-5.
  • 18. D. Mouheb, M. H. Abushamleh, M. H. Abushamleh, Z. A. Aghbari and I. Kamel, “Real- Time Detection of Cyberbullying in Arabic Twitter Streams,” 2019 10th IFIP International Conference on New Technologies, Mobility and Security (NTMS 2019), Canary Islands, Spain, 2019, pp. 1-5.
  • 19. I. Nazar, D. Zois and M. Yao, “A Hierarchical Approach for Timely Cyberbullying Detection,”
  • 2019 IEEE Data Science Workshop (DSW 2019), Minneapolis, MN, 2019, pp. 190-195.
  • 20. H. Nurrahmi and D. Nurjanah, “Indonesian Twitter Cyberbullying Detection Using Text Classification and User Credibility,” 2018 International Conference on Information and Communications Technology (ICOIACT 2018), Yogyakarta, 2018, pp. 543-548.
  • 21. A. Pascucci, V. Masucci and J. Monti, “Computational Stylometry and Machine Learning for Gender and Age Detection in Cyberbullying Texts,” 2019 8th International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW 2019), Cambridge, UK, 2019, pp. 1-6.
 
<<   CONTENTS   >>

Related topics