Discovering Periodic Patterns for Large Scale Mobile Traffic Data: Method and Applications

Shi, Hongzhi; Li, Yong

doi:10.1109/tmc.2018.2799945

Cited by 23 publications

(11 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Several works in the literature have analysed cellular networks traffic traces to understand and model traffic patterns, especially in urban environments [6], [7] . In [6] a heterogeneous dataset containing mobile traffic of ten international cities is investigated with the goal of i) summarising the mobile traffic activity in each area and ii) grouping similar area signatures into a limited representative set.…”

Section: Related Workmentioning

confidence: 99%

Forecasting Busy-Hour Downlink Traffic in Cellular Networks

Pimpinella¹,

Giusto²,

Redondi³

et al. 2022

Preprint

View full text Add to dashboard Cite

The dramatic growth in cellular traffic volume requires cellular network operators to develop strategies to carefully dimension and manage the available network resources. Forecasting traffic volumes is a fundamental building block for any proactive management strategy and is therefore of great interest in such a context. Differently from what found in the literature, where network traffic is generally predicted in the short-term, in this work we tackle the problem of forecasting busy hour traffic, i.e., the time series of observed daily maxima traffic volumes. We tackle specifically forecasting in the long term (one, two months ahead) and we compare different approaches for the task at hand, considering different forecasting algorithms as well as relying or not on a cluster-based approach which first groups network cells with similar busy hour traffic profiles and then fits per-cluster forecasting models to predict the traffic loads. Results on a real cellular network dataset show that busy hour traffic can be forecasted with errors below 10% for look-ahead periods up to 2 months in the future. Moreover, when clusters are available, we improve forecasting accuracy up to 8% and 5% for look-ahead of 1 and 2 months, respectively.

show abstract

Section: Related Workmentioning

confidence: 99%

Forecasting Busy-Hour Downlink Traffic in Cellular Networks

Pimpinella¹,

Giusto²,

Redondi³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…The simple example indicates that the proposed NPMI coefficient can model the correlation well and serve as a preferable metric in the spectrum allocation. There are many discovery schemes for patterns similar with NPMI in the literature [30], [31]. So, we do not probe into the details of the computation of NPMI and focus on its application in the spectrum allocation.…”

Section: Correlation Of Rbs Usability and The Proposed Coefficientmentioning

confidence: 99%

Two-Stage Stable Matching Based Dynamic Spectrum Allocation in Heterogeneous Networks

et al. 2019

View full text Add to dashboard Cite

This paper addresses the resource allocation problem for small cells in the two-tier heterogeneous networks (HetNets) with non-independent resource blocks (RBs) usability, which is prevalent in HetNets application. To formulate the correlations of different RBs to facilitate the resource allocation, a normalized pointwise mutual information (NPMI)-based coefficient is proposed, which is easy to compute and adapt based on the historical information about RBs usability. To obtain a fast allocation solution, a twostage stable matching-based dynamic spectrum allocation algorithm (TSSM) is proposed, which takes both network throughput and user performance as optimization metrics. Specifically, through robust preferable stable matching in the first stage and the iterative matchings with NPMI-based coefficient as the preference profiles in the second stage, the efficient matching solution can be obtained with low complexity and thus is well adapted to the requirement of the resource allocation in HetNets. The validity of the proposed coefficient in modeling the relevance of RBs usability and the benefit of its usage in the allocation algorithm is verified with typical spectrum usability related data. The extensive simulation results on various spectrum environment show the average probability of no RB to use following the proposed TSSM are reduced by more than 13.2%, and the robust requirement for users can be guaranteed better at the same time. Furthermore, the accumulated throughput of all users in the network increases by more than 19.4% on an average under different RBs usability environment. INDEX TERMS Heterogeneous networks, dynamic spectrum allocation, correlation of spectrum usability, two-stage stable matching.

show abstract

“…There are previous studies that investigate mobile user data usage behavior and have focussed on user location and user mobility patterns [14], [15], temporal dynamics [16], and Quality of Experience (QoE) [17], [18], [19]. However, most of the previous studies are either limited to a single operator [20], only target a specific city and location [21], study data usage behaviour targeting application types accessed by users [22], [23], [24], [25], or consider only a few measurement data and user spaces [26]. Unlike these studies, our work uses a large crowd-based dataset collected using Netradar [27] -a mobile network measurement platform.…”

Section: Introductionmentioning

confidence: 99%

Understanding Data Usage Patterns of Geographically Diverse Mobile Users

Walelgne

Asrese

Manner

et al. 2021

IEEE Trans. Netw. Serv. Manage.

View full text Add to dashboard Cite

The increasing trend of the traffic demand from mobile users and the presence of limited resources creates a challenge for network resource management. Understanding the data usage pattern and traffic demand of mobile users is a way forward to enable data-driven network resource management. However, due to the complex nature of mobile networks, understanding and characterizing data usage pattern of mobile users is a daunting task. In this work, we investigate and characterize data usage patterns and behavior of users in mobile networks. We leverage a dataset (∼340M records) collected through a crowd-based mobile network measurement platform -Netradar -across six countries. We elucidate different network factors and study how they affect the data usage patterns by taking mobile users in Finland as a use case. We perform a comparison on data usage patterns of mobile users across six countries by considering total data consumption, network access, the number of sessions created per user, throughput, and user satisfaction level on services. We show that data usage behavior of users over a mobile network is primarily driven by user mobility, the type of data subscription plan marketed by Mobile Network Operators (MNOs), network congestion, and network coverage. Besides, the data usage patterns over different network technologies (e.g., preferring cellular over WiFi) and the percentage of users accessing congested networks vary by country; mostly due to the market pricing strategy and radio coverage. However, the overall data consumption (cellular and WiFi) is comparatively similar in most of the countries we studied.

show abstract

Discovering Periodic Patterns for Large Scale Mobile Traffic Data: Method and Applications

Cited by 23 publications

References 33 publications

Forecasting Busy-Hour Downlink Traffic in Cellular Networks

Forecasting Busy-Hour Downlink Traffic in Cellular Networks

Two-Stage Stable Matching Based Dynamic Spectrum Allocation in Heterogeneous Networks

Understanding Data Usage Patterns of Geographically Diverse Mobile Users

Contact Info

Product

Resources

About