Forecasting user behaviors in call detail records using LSTM models

Download

12626172.pdf

Date

2021-2-13

Author

Kocaman, Hasan

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

534
views

536
downloads

City planners, governors and mobile phone operators utilize quite from Call DetailRecords (CDR) data in the fields like optimizations of traffic congestion, event de-tection, billing and advertisement policies. Both individual and crowd analyses helpimproving the quality of services.In this thesis, we present regression and classification analysis for three main prob-lems. In regression analysis, we experiment on call counts and call time-sums ofspecified numbers of users. We propose cluster based and outlier separating modelsin these two tasks for the purpose of improving the results of individual user-basedmodels comprised of various Long Short Term Memory(LSTM) layers. In the clas-sification analysis, on the other hand, we present models that predict next locationson the trajectories of the users. We improve the results of base LSTM model withtwo-predictions-at-once approach. The analyses show that recurrent neural networkswork well with sequential data and optimizations on top of the models yield promis-ing results.

Subject Keywords

Call detail records, Artificial neural networks, Cluster analysis, Se-quential data

URI

https://hdl.handle.net/11511/89799

Collections

Graduate School of Natural and Applied Sciences, Thesis

Suggestions

OpenMETU
Core

Modelling Green Femtocells in Smart-grids Al-Turjman, Fadi M. (Springer Science and Business Media LLC, 2018-08-01) On going demands to connect massive amounts of the heterogeneous mobile devices and data traffics make the mobile operators in desperate need to find energy-efficient solutions for coverage and real-time services. Accordingly, mobile femtocells are found to be a promising solution in the coming few decades. This paper presents energy-based analysis for mobile femtocells in Ultra-Large Scale (ULS) applications such as the smart-grid. The potential reduction of the consumed energy and service interruption due...
Track ME! A Web Based Location Tracking and Analysis System for Smart Phone Users Bayir, Murat Ali; Demirbas, Murat; Coşar, Ahmet (2009-09-16) Mobility information of cell phone users is very important for wide range of applications, including context-based search and advertising, early warning systems, city-wide sensing applications such as air pollution exposure estimation and traffic planning. With the inclusion of new technologies in the cell phone hardware such as built-in GPS and 802.11 supports, mobility information are easily captured, managed and forwarded to a remote system via opportunistic connections over Internet. However, it is very...
Traffic Event Related Blog Post Classification by Using Traffic Related Named Entities Dundar Unsal, A.; Tüydeş Yaman, Hediye; Karagöz, Pınar (2019-02-28) Real-time monitoring of traffic flow requires physical sensors to be deployed on road networks. Development of such systems might be impractical due to deployment costs of sensors on large scale networks. This study presents a method to extract traffic event related tweets from social streams in order to employ users of social media as human sensors of traffic conditions or events. The proposed method offers a cost effective way of monitoring events or conditions affecting traffic flow. The method consists ...
Adaptable Safety and Security in V2X Systems VİLLARREAL VASQUEZ, Miguel; BHARGAVA, Bharat; Angın, Pelin (2017-06-30) With the advances in the areas of mobile computing and wireless communications, V2X systems have become a promising technology enabling deployment of applications providing road safety, traffic efficiency and infotainment. Due to their increasing popularity, V2X networks have become a major target for attackers, making them vulnerable to security threats and network conditions, and thus affecting the safety of passengers, vehicles and roads. Existing research in V2X does not effectively address the safety, ...
Health monitoring of a bridge system using strong motion data Mosalam, K. M.; Arıcı, Yalın (2009-07-01) In this paper, the acceptability of system identification results for health monitoring of instrumented bridges is addressed. This is conducted by comparing the confidence intervals of identified modal parameters for a bridge in California, namely Truckee 180/Truckee river bridge, with the change of these parameters caused by several damage scenarios. A challenge to the accuracy of the identified modal parameters involves consequences regarding the damage detection and health monitoring, as some of the iden...

Citation Formats

H. Kocaman, “Forecasting user behaviors in call detail records using LSTM models,” M.S. - Master of Science, Middle East Technical University, 2021.