Predicting Customer Quality of Service and Classifying Customer Complaints of a Large Fixed Broadband Service Provider using Machine Learning

21 Mar

Antonio Rocha, Associate Professor at the Institute of Computing at the Fluminense Federal, Brazil

External Presentation (External Speaker)

As in many other organizations, broadband access providers use Active Network Measurements and Trouble Ticket Systems to identify, record and manage problems. However, in large internet access providers, the high number customers bring problems such as, (i) the difficulty to proactively identify the custumers’qualite of service though performing active network measurement; and (ii) given the high amount of complaints, automatically classifying customer complaints reported by trouble ticket systems. In one of my projects, I partner with TIM, one of the largest fixed cell companies and broadband service providers in Brazil, with the main objective of: (i) predicting customers’ Quality of Service (QoS) parameters; and, (ii) automatically classifying customer complaints related to fixed broadband service. To cope with objective (i), we build a framework using Error-Correcting Output Codes (ECOC) and H2O’s Automatic Machine Learning (AutoML) that accurately predicts the quality of service, particularly the download rate, achieved by the customers using features related to customer location, internet plan, and equipment. Our experiments demonstrate that our model achieves around 83% accuracy on average on our dataset. Our framework can be used by TIM to improve their fixed broadband services. To cope with objective (ii) we propose a methodology to automate the process of allocating a trouble ticket, registered in a call center, to the technical team with the necessary knowledge to solve it. Through a custom data preprocessing in conjunction with the application of Machine Learning algorithms, this work achieves accuracy of 89%, outperforming several similar works. Our work can assist TIM to improve their complaint resolution process. At the end of this talk, I will present some possible opportunities of collaborations with other research groups in this and other areas of interest.

About Antonio Rocha

Antonio Rocha is Associate Professor in the Computer Science Department from the Institute of Computing at the Fluminense Federal since 2011. He received his MSc and PhD degrees in Computer and Systems Engineering (PESC/COPPE) from the Federal University of Rio de Janeiro (UFRJ) Brazil, in 2003 and 2010, respectively. During PhD, in 2008-2009, he was a visiting student in Computer Science at University of Massachusetts-Amherst (UMass). In 2010, he worked as a post-doc researcher at UFRJ, supported by INCT WebScience. Recently, he returned for a sabbatical as a visiting professor at University of Massachusetts-Amherst. He is awarded as Research Productivity Fellowship granted by CNPq and Young Scientist of Rio de Janeiro by FAPERJ. His areas of interest include performance evaluation, traffic engineering, network measurement, next generation Internet, network science and security systems. Dr. Antonio Rocha has published many papers in important journals and conferences, and some of those works received a few awards, such as Best Papers in ACM/CoNEXT, SBC/SBRC and SBC/WPerformance, and nominated among the top-6 PhD theses from Computer Brazilian Society in 2012.

This event will be conducted in English

  • Location: MR-A1 [Ramón] & MR-A2 [Cajal], IMDEA Networks Institute, Avda. del Mar Mediterráneo 22, 28918 Leganés – Madrid
  • Organization: NETCOM Research Group (Telematics Engineering Department, UC3M); IMDEA Networks Institute
  • Time: 14:00
  • Add to Calendar: iCalendar Outlook Google