Radiotekhnika
Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS


Тел.: +7 (495) 625-9241

 

Analysis of some types and architectures of neural networks for predicting the wait time of the income call of the call center

Keywords:

Yu.V. Koltsov – Ph.D. (Phys.-Math.), Associate Professor, Kuban State University
N.G. Repkin – Scientist, Bank «Pervomajskij»


Contact centre is the subdivision within any organization which applies specific technologies and equipment for verification of the great amount of phone calls during which the collection of the statistic data is implemented along with the differentiation of the control and monitoring means. One of the key aspects of the call centre effectiveness is the average time expecting tine of the call in the line. The count of the expectation time for the incoming call allows choosing the best way to verify the calls and automatically inform the subscriber how much time they need to wait in the line. However the forecasting of the expecting time in the line is too difficult because you need to take into consideration the huge amount of the constantly changing factors. We offer the method based on the calculation of the above-described aspect by the few neural nets in the following way: cascade of neural nets accepts some aspects connected with the work of the call centre (operational and statistical) and the cluster net keeps the income call in accordance with one of the classes defined in advance. Each if these classes has the preliminary educated net structure which is able to calculate the approximate expecting time in accordance with data transferred into it. The article also describes several architectures of the neuron nets appropriate for the solution of the described task as well as the algorithms of nets education. Beside we have also revealed the indicators of the contact centre work which are also the most appropriate for the classification of the income calls.
References:

  1. Kol'czov Ju.V., Repkin N.G. Nejrosetevaya model' prognozirovaniya vremeni ozhidaniya zayavki SMO // Matematika programmny'x sistem. 2010. № 7. S. 49–54.
  2. Ossovskij S. Nejronny'e seti dlya obrabotki informaczii / Per. s pol'skogo I.D. Rudinskogo. M.: Finansy' i statistika. 2002.
  3. Xajkin S. Nejronny'e seti. Polny'j kurs. Izd. 2-e / Per. s angl. M.: Vil'yams. 2008.
  4. Savel'ev A.V. Obshhaya teoriya samoorganizaczionnogo nejroupravleniya // Nejrokomp'yutery': razrabotka, primenenie. 2013. № 5. S. 3-13.
  5. E'lektronny'j uchebnik StatSoft. Mnogoslojny'j perseptron http://www.statsoft.ru/home/portal/applications/neuralnet­worksadvisor/adv-new/MultilayerPerceptrons.htm.

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio