Revealing Infrastructure-Stressing Clients in the Customer Base of a Scandinavian Operator using and HPI Future SoC Lab Hardware Resources
2019 (Engelska) Ingår i: Technische Berichte des Hasso-Plattner-Instituts fur Softwaresystemtechnik an der Universitat Potsdam, Universitatsverlag Potsdam , 2019, Vol. 130, s. 59-65Konferensbidrag, Publicerat paper (Refereegranskat)
Abstract [en]
We define the term an Infrastructure-Stressing client. Roughly speaking, she uses the infrastructure in a taxing manner, such as always staying in the zones of high demand. We developed a method based on combinatorial optimization to reveal the Infrastructure-Stressing clients in the customer base based on trajectory information from Call Data Records. We have found that 7 % in the customer base are Infrastructure-Stressing. As was expected, a correlation exists between this quality and client's geo-demographic segment. Currently we are working on a predictive model to be able to tell an Infrastructure-Stressing client in a newcomer whose mobility is yet unknown to the operator. © 2019 Universitatsverlag Potsdam. All rights reserved.
Ort, förlag, år, upplaga, sidor Universitatsverlag Potsdam , 2019. Vol. 130, s. 59-65
Nyckelord [en]
Sales, System-on-chip, Call data records, Customerbase, Hardware resources, High demand, Predictive models, Trajectory information, Combinatorial optimization
Nationell ämneskategori
Datavetenskap (datalogi)
Identifikatorer URN: urn:nbn:se:bth-26942 Scopus ID: 2-s2.0-85160719856 ISBN: 9783869564753 (tryckt) OAI: oai:DiVA.org:bth-26942 DiVA, id: diva2:1900605
Konferens HPI Future SOC Lab 2017, Potsdam, Nov 15 2017
2024-09-242024-09-242025-09-30 Bibliografiskt granskad