Prediction of Telecom Churns and Consumer Behaviour using Recurrent Neural Networks
Lakshay Arora1, Aayush Kapur2, Lavanya K3
1Lakshay Arora, VIT University, Vellore (Tamil Nadu), India.
2Aayush Kapur, VIT University, Vellore (Tamil Nadu), India.
3Lavanya K, VIT University, Vellore (Tamil Nadu), India.
Manuscript received on 18 April 2019 | Revised Manuscript received on 25 April 2019 | Manuscript published on 30 April 2019 | PP: 353-358 | Volume-8 Issue-4, April 2019 | Retrieval Number: D6142048419/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: In today’s world, telecommunication is a very common thing. People register on a company’s platform. However, due to various issues a normal person might churns or unsubscribe to a particular telecommunication provider. So being able to predict churn is a pretty handy tool to a marketer or advertiser. They would be able to predict if a user will probably quit their telecommunication network or not. In this paper, we created a telecommunication churn prediction system by using recurrent neural networks. The basic idea of this is to create a system so that we can predict consumer behaviour and tell us whether a consumer will want to give a service based on his/her calling patterns, recharge frequency and amount and a host of other factors. This has been achieved by the help of recurrent neural networks. Recurrent neural networks are basically normal neural networks with a feedback loop. It takes its generic input and the output of the previous stage in each neuron. We initially, pre-processed the data, then creating the model of RNN by varying different parameters. The data was then passed to the RNN to train the model. Once that was done, the model was tested with testing data and results were produced. We were able to get pretty good accuracy with our model.
Keywords: Churn Prediction, Feedback Loop, Neural Network, RNN
Scope of the Article: Neural Information Processing