Yao / Su / Tong Mobile Data Mining

1. Auflage 2018
ISBN: 978-3-030-02101-6
Verlag: Springer International Publishing
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 64 Seiten

Reihe: SpringerBriefs in Computer Science

This SpringerBrief presents a typical life-cycle of mobile data mining applications, including:data capturing and processing which determines what data to collect, how to collect these data, and how to reduce the noise in the data based on smartphone sensors
 feature engineering which extracts and selects features to serve as the input of algorithms based on the collected and processed data
 model and algorithm designIn particular, this brief concentrates on the model and algorithm design aspect, and explains three challenging requirements of mobile data mining applications: energy-saving, personalization, and real-time
 Energy saving is a fundamental requirement of mobile applications, due to the limited battery capacity of smartphones. The authors  explore the existing practices in the methodology level (e.g. by designing hierarchical models) for saving energy. Another fundamental requirement of mobile applications is personalization.  Most of the existing methods tend to train generic models for all users, but the authors provide existing personalized treatments for mobile applications, as the behaviors may differ greatly from one user to another in many mobile applications. The third requirement is real-time. That is, the mobile application should return responses in a real-time manner, meanwhile balancing effectiveness and efficiency. This SpringerBrief targets data mining and machine learning researchers and practitioners working in these related fields. Advanced level students studying computer science and electrical engineering will also find this brief useful as a study guide. 

Autoren/Hrsg.

Yao, Yuan

Su, Xing

Tong, Hanghang

Inhaltsverzeichnis

1;Preface;6
2;Contents;7
3;Acronyms;9
4;1 Introduction;10
4.1;1.1 Background;10
4.2;1.2 Typical Applications;11
4.3;1.3 Steps, Characteristics, and Challenges;12
4.4;1.4 Roadmap;14
5;2 Data Capturing and Processing;16
5.1;2.1 Smartphone Sensors;16
5.2;2.2 Data Collection;17
5.3;2.3 Data Denoising;21
5.4;2.4 Summary;25
6;3 Feature Engineering;26
6.1;3.1 Data Segmentation;26
6.2;3.2 Feature Extraction;27
6.3;3.3 Feature Analysis and Sensor Selection;28
6.4;3.4 Summary;32
7;4 Hierarchical Model;33
7.1;4.1 Problem Description;33
7.2;4.2 A Hierarchical Framework;34
7.3;4.3 Experimental Evaluations;36
7.4;4.4 Summary;38
8;5 Personalized Model;39
8.1;5.1 Problem Description;39
8.2;5.2 The Personalized Approach: Overview;40
8.3;5.3 The Personalized Approach: Details;42
8.3.1;5.3.1 Similarity Computation;42
8.3.2;5.3.2 Distribution Estimation;43
8.3.3;5.3.3 Sample Selection;44
8.3.4;5.3.4 Sample Reweighting;44
8.3.5;5.3.5 Algorithm Analysis;45
8.4;5.4 Experimental Evaluations;46
8.4.1;5.4.1 Experiment Setup;47
8.4.2;5.4.2 Experiment Results;47
8.5;5.5 Summary;49
9;6 Online Model;50
9.1;6.1 Problem Description;50
9.2;6.2 Online Learning;51
9.3;6.3 Experimental Evaluations;52
9.3.1;6.3.1 Online Learning vs. Offline Learning;53
9.3.2;6.3.2 How Much Data Are Enough to Train an Initial Model?;56
9.3.3;6.3.3 Combining with the Hierarchical Model;56
9.4;6.4 Summary;57
10;7 Conclusions;58
10.1;7.1 Summary;58
10.2;7.2 Discussions;59
10.2.1;7.2.1 More Combinations of Sensors;59
10.2.2;7.2.2 More Usages of Smartphone Sensors;60
11;References;61