Caller behaviour classification using computational intelligence methods

Pretesh B. Patel, Tshilidzi Marwala

Research output: Contribution to journalArticlepeer-review

10 Citations (Scopus)


A classification system that accurately categorizes caller interaction within Interactive Voice Response systems is essential in determining caller behaviour. Field and call performance classifier for pay beneficiary application are developed. Genetic Algorithms, Multi-Layer Perceptron neural network, Radial Basis Function neural network, Fuzzy Inference Systems and Support Vector Machine computational intelligent techniques were considered in this research. Exceptional results were achieved. Classifiers with accuracy values greater than 90% were developed. The preferred models for field 'Say amount', 'Say confirmation' and call performance classification are the ensemble of classifiers. However, the Multi-Layer Perceptron classifiers performed the best in field 'Say account' and 'Select beneficiary' classification.

Original languageEnglish
Pages (from-to)87-93
Number of pages7
JournalInternational Journal of Neural Systems
Issue number1
Publication statusPublished - Feb 2010


  • Caller behaviour
  • Fuzzy inference system
  • Interactive voice response
  • Multi-layer perceptron
  • Radial basis function
  • Support vector machines

ASJC Scopus subject areas

  • Computer Networks and Communications


Dive into the research topics of 'Caller behaviour classification using computational intelligence methods'. Together they form a unique fingerprint.

Cite this