Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/107047
Title: Modelización de temas de llamadas en tiempo real
Author: Gómez Montero, Manuel Enrique
Director: Casas-Roma, Jordi  
Tutor: Valdivia García, Ana
Abstract: A call center is the area of a company in charge of receiving and transmitting calls to or from customers, business partners or other external companies. The vast amount of information generated from these calls reflects, among other aspects, the needs of the company's customers. The ability to classify the calls received in real time according to the subject, would lead to an improvement in the quality of the company's customer service as well as the customer experience. This is approachable through machine learning techniques. This draft reveals a system implemented in the call center of Telefónica that allows us, through the transcription of the calls, to know the topic of the calls using Natural Language Processing methods. The analysis mentioned above makes classifications with different architectures, with unsupervised models as well as deep learning supervised models. Depending on the model, different methods have also been studied when representing the transcripts. Finally, one of these models has been moved to production, giving the system the possibility of making this classification in real time and providing the company the ability of visualizing the evolution of the calls over time and generating alerts in the event of any anomaly. The system moved to production has been fully deployed in containers and equipped with monitoring mechanisms, continuous integration and continuous deployment. The project opens up a vast number of possibilities for the company to carry out new Natural Language Processing systems. Also, to bring deep learning models to a productive environment.
Keywords: real time
call center
natural language processing
sentiment analysis
deep learning
topic modeling
Document type: info:eu-repo/semantics/masterThesis
Issue Date: 25-Jan-2019
Publication license: http://creativecommons.org/licenses/by-sa/3.0/es/  
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
mnlgmonteroTFM0120memoria.pdfMemoria del TFM12,9 MBAdobe PDFThumbnail
View/Open