Performance of svm, k-nn and nbc classifiers for text-independent speaker identification with and without modelling through merging models

Abstract

This paper proposes some methods of robust text-independent speaker identification based on Gaussian Mixture Model (GMM). We implemented a combination of GMM model with a set of classifiers such as Support Vector Machine (SVM), K-Nearest Neighbour (K-NN), and Naive Bayes Classifier (NBC). In order to improve the identification rate, we developed a combination of hybrid systems by using validation technique. The experiments were performed on the dialect DR1 of the TIMIT corpus. The results have showed a better performance for the developed technique compared to the individual techniques.

Keywords

• GMM
• SVM
• KNN
• NB
• TIMIT

Birleşik modellemeli ve modellemesiz metin-bağımsız konuşmacı tanıma için SVM, K-NN ve NBC sınıflandırıcıların başarımı

Abstract

Bu çalışma Gaussian Mixture Model tabanlı metin-bağımsız konuşmacı tanıma yöntemleri sunar. GMM model ile Support Vector Machine, K-nearest Neighbour ve Naive Bayes sınıflandırıcı gibi sınıflandırıcıların kombinasyonu gerçekleştirilmiştir. Tanıma oranını iyileştirmek için, doğrulama yöntemi kullanarak hibrid sistemlerin kombinasyonunu geliştirdik. Deneyler TIMIT corpus’ un DR1 lehçesi üzerine yapılmıştır. Sonuçlar ayrı ayrı yöntemlerle karılaştırıldığında geliştirilen yöntemle daha iyi başarım göstermiştir.

Keywords

• GMM
• Combination
• SVM
• KNN
• NB
• TIMIT

References

1. D. A. N. R.Amami, “An Empirical Comparison
2. of SVM and Some Supervised Learning
3. Algorithms for Vowel recognition”,
4. International Journal of Intelligent Information
5. Processing IJIIP, 2012.
6. B.S. Atal, “Automatic Recognition of Speaker
7. from Their Voices”, Proceedings of the IEEE,
8. Vol. 64, No. 4, pp 460-475, 1976

9. W. M. Campbell, D. E. Sturim, D. A. Reynolds,
10. and A. Solomon off, “SVM based speaker
11. verification using a GMM supervector kernel and
12. Without modelling GMM
13. Classifiers SVM K-NN NB
14. Identification rate
15. (%)
16. 87 92
17. Y. Nahayo, S. Arı Performance of svm, k-nn and nbc classifiers for text
18. independent speaker identification with and without
19. modelling through merging models
20. SAÜ Fen Bil Der 20. Cilt, 1. Sayı, s. 1-6, 2016
21. NAP variability compensation”, Proc. Int. Conf.
22. Acoustics, Speech, and Signal Processing, 2006.
23. D. Reynolds and R. Rose, "Robust textindependent speaker identification using
24. Gaussian mixture speaker models, " IEEE Trans.
25. Speech Audio Proc., vol. 3, no. 1, pp. 72–83,
26. -
27. D. Ben Ayed Mezghani, S. Zribi Boujelbene et
28. N. Ellouze, "Evaluation of SVM kernels function
29. and conventional machine learning algorithms
30. for Speaker Identification Task," International
31. Journal of Hybrid Information Technology
32. (IJHIT), vol. 3, pp. 2 3-34, 2010.
33. S. Zribi Boujelbene, D. Ben Ayed Mezghan et N.
34. Ellouze, "Application of Combining Classifiers
35. for Text-Independent Speaker Identification,"
36. the 16th IEEE International Conference on
37. Electronics, Circuits, and Systems ICECS,
38. Hammamet-Tunisie, pp. 723-726, 2009.
39. R. Djemili, M. Bedda and H. Bourouba, "A
40. Hybrid GMM/SVM System for Text
41. Independent Speaker Identification,"
42. International Journal of Computer and
43. Information Science and Engineering, vol. 1, pp.
44. -8, 2007.
45. D. Neiberg “Text Independent Speaker
46. Verification Using Adapted Gaussian Mixture
47. Models", Centre for Speech Technology (CTT)
48. Department of Speech, Music and Hearing KTH,
49. Stockholm, Sweden 2001-12-11.
50. S. Zribi Boujelbene, D. Ben Ayed Mezghan et N.
51. Ellouze, “Support Vector Machines approaches
52. and its application to speaker identification," 3rd
53. IEEE International Conference on Digital
54. Ecosystems and Technologies DEST, pp. 662-
55. , 2009.
56. I. Ayed, “Stratégies de fusion de paramètres pour
57. une tâche d'identification du locuteur en mode
58. indépendant du texte : Application sur le corpus
59. NTIMIT.TAIMA”, Hammamet-Tunisie 2011.
60. L. Lam et C.Y. Suen. “Application of Majority
61. Voting to Pattern Recognition: An Analysis of Its
62. Behavior and Performance”, IEEE Transactions
63. on Systems, Man Cybernetics, pp. 553-568,
64. -
65. K. S. Durgesh, and B. Lekha, “Data classification
66. using support vector machine.” Journal of
67. Theoretical and Applied Information
68. Technology, 12(1), 1-7, 2010
69. H. Y. Chang, A. L. Kong, and L. Haizhou, “An
70. SVM Kernel With GMM-Supervector Based on
71. the Bhattacharyya Distance for Speaker
72. Recognition”, v.6, pp. 1300-1312, 2010