Multiclass Digital Audio Segmentation with MFCC Features using Naive Bayes and SVM Classifiers

Iheme, Leonardo O.; Ozan, Sukru

doi:10.1109/asyu48272.2019.8946441

Cited by 5 publications

(7 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, details regarding the number and type of datasets are missing and only a single classification factor is discussed. In [81], the authors proposed audio segments classification data applied to call centers using Naïve Bayes and SVM. Experiments have shown that SVM achieved 83% accuracy when MFCC is used with first and second derivatives and further 87% is achieved when Naïve Bayes is used.…”

Section: B Classification Based Approaches Used For Load Balancingmentioning

confidence: 99%

A Hybrid Model for Load Balancing in Cloud Using File Type Formatting

et al. 2020

View full text Add to dashboard Cite

Maintaining accuracy in load balancing using metaheuristics is a difficult task even with the help of recent hybrid approaches. In the existing literature, various optimized metaheuristic approaches are being used to achieve their combined benefits for proper load balancing in the cloud. These approaches often adopt multi-objective QoS metrics, such as reduced SLA violations, reduced makespan, high throughput, low overload, low energy consumption, high optimization, minimum migrations, and higher response time. The cloud applications are generally computation-intensive and can grow exponentially in memory with the increase in size if no proper effective and efficient load balancing technique is adopted resulting in poor quality solutions. To provide a better load balancing solution in cloud computing, with extensive data, a new hybrid model is being proposed that performs classification on the number of files present in the cloud using file type formatting. The classification is performed using Support Vector Machine (SVM) considering various file formats such as audio, video, text maps, and images in the cloud. The resultant data class provides high classification accuracy which is further fed into a metaheuristic algorithm namely Ant Colony Optimization (ACO) using File Type Formatting FTF for better load balancing in the cloud. Frequently used QoS metrics, such as SLA violations, migration time, throughput time, overhead time, and optimization time are evaluated in the cloud environment and comparative analysis is performed with recent metaheuristics, such as Ant Colony Optimization-Particle Swarm Optimization (ACOPS), Chaotic Particle Swarm Optimization (CPSO), Q-learning Modified Particle Swarm Optimization (QMPSO), Cat Swarm Optimization (CSO) and D-ACOELB. The proposed algorithm outperforms them and provides good performance with scalability and robustness.

show abstract

Section: B Classification Based Approaches Used For Load Balancingmentioning

confidence: 99%

A Hybrid Model for Load Balancing in Cloud Using File Type Formatting

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Classification of multimedia data needs preprocessing phase which performs complex task to identify the classes. Navie bayes algorithm provides the chances of occurrences of the particular data related to problem area, but it requires other algorithm support to improve accuracy [22]. Credit score are calculated with respective models performs the operation in two ways online and offline.…”

Section: Fig 5histogram Analysis 5 Results and Discussionmentioning

confidence: 99%

An Efficient Machine Learning Model for Location Aware Credit Fraud and Risk Classification and Detection

Muthulakshmi¹,

Saravanakumar²,

Tamizhselvi³

2021

Proceedings of the Fist International Conference on Advanced Scientific Innovation in Science, Engineering and Technology, ICA

View full text Add to dashboard Cite

Online transaction grows in enormous rate because of the strength of usage by the user. User always use online mode to pay the amount to the respective merchant. Various method of payment is available in the market but credit card is so popular due to the pre credit is assigned to the customer by banker. Card user gets extra time for paying the payment which gives comfortable live to them. Security of the card suffers in various factors such as theft, fraud, illegal access, so it is protected by using modern algorithm with automated capability. Artificial Intelligent algorithms are applied to detect the fraud but that is not achieving enough accuracy. This type of problem is overcome by using location based risk identification model with multidimensional features for analysis. Three phases of processing is carried out namely feature management, risk management and Location awareness. The focus of the model is to protect the credit card frauds in multi level security by identifying the source and location of access. It achieves high level of security when compared to all exiting algorithms with reliable manner.

show abstract

“…Ses işleme araştırmalarında MFCC önemli bir yere sahip olup yüksek başarı oranıyla ses verisinden öznitelik elde edilmesinde yaygın olarak kullanılan bir tekniktir [4]. Ses verilerinden çıkarılan MFCC'nin makine öğrenmesi ve derin öğrenme algoritmalarında öznitelik olarak kullanılmasıyla dijital ses işleme alanında birçok çalışma gerçekleştirilmiştir [5][6][7][8][9][10][11]. Turnbull vd.…”

Section: Gi̇ri̇ş (Introduction)unclassified

“…Iheme vd. ses verisinin sessizlik, konuşma ve müzik olmak üzere üç sınıftan biri olarak sınıflandırmak amacıyla ses verilerinin MFCC katsayılarını, türevlerini ve ikinci türevlerini SVM (Support Vector Machine) ve Naive Bayes algoritmalarının eğitiminde öznitelik olarak kullanmış ve bu özniteliklerin başarıya etkilerini araştırmışlardır [6]. MFCC öznitelikleri konuşma duygusu tanımada yaygın olarak kullanılmaktadır.…”

Section: Gi̇ri̇ş (Introduction)unclassified

Detection of Negative Calls in Call Centers with Convolutional Neural Networks

KARATAŞ>¹,

MERCAN²,

ÖZDİL>³

et al. 2023

Bilişim Teknolojileri Dergisi

View full text Add to dashboard Cite

Bu çalışmada çağrı merkezi çalışanları ile müşteriler arasındaki telefon konuşmalarının otomatik olarak olumlu veya olumsuz şeklinde değerlendirilmesi üzerine odaklanılmıştır. Çalışmada kullanılan veri seti firma bünyesinde gerçekleştirilen telefon görüşmelerinden oluşmaktadır. Veri seti üçer saniyelik 10411 adet ses kaydını içermekte olup bu kayıtların 5408 tanesi olumlu kayıtlardan 5003 tanesi münakaşa, öfke ve hakaret içeren olumsuz kayıtlardan oluşmaktadır. Çağrı merkezi kayıtlarından duygu tanıma için anlamlı öznitelikler elde etmek amacıyla her bir ses kaydından MFCC öznitelikleri çıkarılmıştır. Çağrı merkezi kayıtlarını olumlu olumsuz olarak sınıflandırmak için önerilen CNN mimarisi MFCC öznitelikleriyle eğitilmiştir. Önerilen CNN modeli %86,1 eğitim başarısı, %77,3 doğrulama başarısı göstermiş olup test verileri üzerinde %69,4 sınıflandırma başarısı elde edilmiştir. Bu çalışma ile çağrı merkezlerinde gerçekleşen konuşmaların otomatik analizi yapılıp olumsuz durumların kalite yöneticilerine bildirilmesiyle gerekli önlemlerin alınarak müşteri memnuniyetinin artırılması amaçlanmaktadır.

show abstract

Multiclass Digital Audio Segmentation with MFCC Features using Naive Bayes and SVM Classifiers

Cited by 5 publications

References 6 publications

A Hybrid Model for Load Balancing in Cloud Using File Type Formatting

A Hybrid Model for Load Balancing in Cloud Using File Type Formatting

An Efficient Machine Learning Model for Location Aware Credit Fraud and Risk Classification and Detection

Detection of Negative Calls in Call Centers with Convolutional Neural Networks

Contact Info

Product

Resources

About