Premium
Efficient mobility prediction scheme for pervasive networks
Author(s) -
Garg Neeraj,
Dhurandher Sanjay K.,
Nicopolitidis Petros,
Lather J. S.
Publication year - 2018
Publication title -
international journal of communication systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.344
H-Index - 49
eISSN - 1099-1131
pISSN - 1074-5351
DOI - 10.1002/dac.3520
Subject(s) - computer science , probabilistic logic , context (archaeology) , scheme (mathematics) , ubiquitous computing , sequence (biology) , dynamic bayesian network , machine learning , movement (music) , bayesian network , artificial intelligence , chart , data mining , human–computer interaction , mathematical analysis , paleontology , philosophy , statistics , genetics , mathematics , biology , aesthetics
Summary This paper aims towards probabilistic reasoning and Bayesian‐based recommendations to predict the next movement of a person. The proposed model in this work observes the behavior and movement patterns of humans for a day both at home and at their office to predict their future activities. To achieve this, an efficient model has been designed that provides the probable context‐based location of a person and predicts his next movement based on his behavior on some particular day at a particular time. The proposed model allows ubiquitous services to adapt to uncertain situations in today's world using different mechanisms such as monitoring the human behavior patterns and evaluating the user preferences and profiles. A case study of the office activity chart has been provided, and based on the experimentation performed on the related events, the probability in evaluating some “N”chained events of a person in a consecutive order using the proposed model has been found to be 0.002, which infers that there are fewer chances that the person will perform the same particular sequence of events.