Characterization of the Mobile User Profile Based on Sentiments and Network Usage Attributes

被引:0
作者
de Morais, Leonardo P. [1 ,3 ]
Immich, Roger [2 ]
Silva, Nadia Felix [1 ]
Rosa, Thierson Couto [1 ]
Borges, Vinicius da Cunha Martins [1 ]
机构
[1] Univ Fed Goias, Goiania, Brazil
[2] Univ Fed Rio Grande do Norte, Rio Grande do Norte, Brazil
[3] Univ Fed Goias, UFG Alameda Palmeiras, Inst Informat Goiania, Campus Samambaia, BR-74001970 Goiania, GO, Brazil
基金
巴西圣保罗研究基金会;
关键词
Future Mobile Networks; Sentiment Analysis; User Profile; Association Rules; Frequent Item -set Mining; TECHNOLOGIES; CHALLENGES; MANAGEMENT; FRAMEWORK; ISSUES; SDN;
D O I
10.5753/jisa.2022.2520
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Providing resources to meet user needs in futuristic mobile networks is still challenging since the network resources like spectrum and base stations do not increase in the same proportion as the accelerated growth of net-work traffic. Because of this, human/user behavior attributes can assist resource management in dealing with these challenges, which pick up aspects of how the user impacts the usage of mobile networks, such as network usage, the content of interest, urban mobility routines, social networks, and sentiment. A user profile is a combination of user/human behavior attributes. Such profiles are expected to be a knowledge for softwarization enablers to im-prove the management of future wireless networks fully. Nevertheless, the correlation between human sentiment and wireless and mobile network usage has not been deeply investigated in the literature about the mobile user profile. This work aims to define the user profile using a transfer learning approach for the sentiment classification of WhatsApp messages. A real-life experiment was conducted to collect users' attributes, namely the WhatsApp messages and network usage. A new data analysis methodology is proposed that consists of a frequent item-set pattern mining (FP-Growth) based on Association Rules, the Chi-squared statistical test, and descriptive statistics. This methodology assesses the correlation between sentiment and network usage in a profound way. Results show that the users participating in the experiment form three groups. The first group, with 55.6% of the users, contains users who present a strong relation between negative sentiment and low network usage and also a strong relation between positive sentiment and high network usage. The second group contains 25.9% of the users and is composed of users who present a strong relation between positive sentiment and high network usage. The third group contains 18.5% of the users for whom the correlation between sentiment and network usage is still statistical significant, but the strength of this relation is much more weak then in the other two groups. Thus, 81.5% of the users (the first two groups) present a strong relation between user sentiment captured from WhatsApp messages and the network traffic generated by them.
引用
收藏
页码:82 / 97
页数:16
相关论文
共 50 条
  • [31] Research on personalized search engine based on user profile
    Wu, Xiao
    Wang, Peng
    Li, Dan-Ning
    Li, Jian-Ping
    [J]. PROCEEDINGS OF THE INTERNATIONAL CONFERENCE INFORMATION COMPUTING AND AUTOMATION, VOLS 1-3, 2008, : 107 - +
  • [32] Research on Personalized Recommendation Algorithms Based on User Profile
    Guo Hui
    Chen Mang
    Zhou LiQing
    Xv ShiKun
    [J]. INTERNATIONAL JOURNAL OF ADVANCED COMPUTER SCIENCE AND APPLICATIONS, 2024, 15 (03) : 289 - 298
  • [33] A User Profile based Pension Service Recommendation Algorithm
    Li, Chunshan
    Chu, Dianhui
    Xu, Xiaofei
    Bu, Yunfei
    [J]. 2019 IEEE WORLD CONGRESS ON SERVICES (IEEE SERVICES 2019), 2019, : 374 - 375
  • [34] The Prediction of Web User's Behavior Based on Web Usage Mining
    Bai Xiaoli
    Chen Fei
    [J]. 2010 ETP/IITA CONFERENCE ON TELECOMMUNICATION AND INFORMATION (TEIN 2010), 2010, : 32 - 35
  • [35] Simulation of Trust-Based Mechanism for Enhancing User Confidence in Mobile Crowdsensing Systems
    Kalui, Dorothy Mwongeli
    Zhang, Dezheng
    Muketha, Geoffrey Muchiri
    Onsomu, Jared Okoyo
    [J]. IEEE ACCESS, 2020, 8 (08): : 20870 - 20883
  • [36] User-Oriented Virtual Mobile Network Resource Management for Vehicle Communications
    Lu, Huimin
    Zhang, Yin
    Li, Yujie
    Jiang, Chi
    Abbas, Haider
    [J]. IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, 2021, 22 (06) : 3521 - 3532
  • [37] Offloading Personal Security Applications to the network edge: a Mobile User case scenario
    Montero, D.
    Serral-Gracia, R.
    [J]. 2016 INTERNATIONAL WIRELESS COMMUNICATIONS AND MOBILE COMPUTING CONFERENCE (IWCMC), 2016, : 96 - 101
  • [38] A Method for User Profile Learning in Document Retrieval System Using Bayesian Network
    Maleszka, Bernadetta
    [J]. INTELLIGENT INFORMATION AND DATABASE SYSTEMS, ACIIDS 2017, PT I, 2017, 10191 : 269 - 277
  • [39] User Cooperation for NOMA-Based Mobile Edge Computing
    Huang, Yuwen
    Liu, Yuan
    [J]. PROCEEDINGS OF 2018 IEEE INTERNATIONAL CONFERENCE ON COMMUNICATION SYSTEMS (ICCS 2018), 2018, : 395 - 400
  • [40] Exploitation of User Profile in the Information Retrieval based on Bayesian Approach
    Achemoukh, Farida
    Ahmed-Ouamer, Rachid
    [J]. 2013 3RD INTERNATIONAL SYMPOSIUM ISKO-MAGHREB, 2013,