A location based sequence prediction algorithm for determining next activity in smart home

Mohammad Marufuzzaman, Md. Mamun Ibne Reaz, Labonnah Farzana Rahman, Araf Farayez

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Smart home or home automation has become widely popular especially in the case of easing the lives of people with special needs, for instance the elderly and handicapped people. In every home, a specific user has a unique pattern or sequence of using the functions of that house. Recognizing that unique pattern is the key to ensuring an intelligently and properly automated household where the house will remember the behavior of a user and predict the next service required by the user successfully. In this research, a recently developed algorithm named as sequence prediction via enhanced episode discovery (SPEED) is considered for modification by inclusion of location agents. A smart home prototype consisting of two rooms is designed as a testbed for verification. The results show that the accuracy of this algorithm is more than 40%, which is better than the previous SPEED. Moreover, the algorithm detects the location of next predicted event. Since human activity can be distinguished by their existing locations, predicting the next event as well as the location helps to determine the next action more accurately.

Original languageEnglish
Pages (from-to)161-165
Number of pages5
JournalJournal of Engineering Science and Technology Review
Volume10
Issue number2
Publication statusPublished - 2017

Fingerprint

Testbeds
Automation

Keywords

  • Artificial intelligence
  • Hidden markov models
  • Location agent
  • Prediction by partial matching
  • Smart homes

ASJC Scopus subject areas

  • Engineering(all)

Cite this

A location based sequence prediction algorithm for determining next activity in smart home. / Marufuzzaman, Mohammad; Ibne Reaz, Md. Mamun; Rahman, Labonnah Farzana; Farayez, Araf.

In: Journal of Engineering Science and Technology Review, Vol. 10, No. 2, 2017, p. 161-165.

Research output: Contribution to journalArticle

Marufuzzaman, Mohammad ; Ibne Reaz, Md. Mamun ; Rahman, Labonnah Farzana ; Farayez, Araf. / A location based sequence prediction algorithm for determining next activity in smart home. In: Journal of Engineering Science and Technology Review. 2017 ; Vol. 10, No. 2. pp. 161-165.
@article{c9420e5521a54c1cb0d121086e83a375,
title = "A location based sequence prediction algorithm for determining next activity in smart home",
abstract = "Smart home or home automation has become widely popular especially in the case of easing the lives of people with special needs, for instance the elderly and handicapped people. In every home, a specific user has a unique pattern or sequence of using the functions of that house. Recognizing that unique pattern is the key to ensuring an intelligently and properly automated household where the house will remember the behavior of a user and predict the next service required by the user successfully. In this research, a recently developed algorithm named as sequence prediction via enhanced episode discovery (SPEED) is considered for modification by inclusion of location agents. A smart home prototype consisting of two rooms is designed as a testbed for verification. The results show that the accuracy of this algorithm is more than 40{\%}, which is better than the previous SPEED. Moreover, the algorithm detects the location of next predicted event. Since human activity can be distinguished by their existing locations, predicting the next event as well as the location helps to determine the next action more accurately.",
keywords = "Artificial intelligence, Hidden markov models, Location agent, Prediction by partial matching, Smart homes",
author = "Mohammad Marufuzzaman and {Ibne Reaz}, {Md. Mamun} and Rahman, {Labonnah Farzana} and Araf Farayez",
year = "2017",
language = "English",
volume = "10",
pages = "161--165",
journal = "Journal of Engineering Science and Technology Review",
issn = "1791-9320",
publisher = "Kavala Institute of Technology",
number = "2",

}

TY - JOUR

T1 - A location based sequence prediction algorithm for determining next activity in smart home

AU - Marufuzzaman, Mohammad

AU - Ibne Reaz, Md. Mamun

AU - Rahman, Labonnah Farzana

AU - Farayez, Araf

PY - 2017

Y1 - 2017

N2 - Smart home or home automation has become widely popular especially in the case of easing the lives of people with special needs, for instance the elderly and handicapped people. In every home, a specific user has a unique pattern or sequence of using the functions of that house. Recognizing that unique pattern is the key to ensuring an intelligently and properly automated household where the house will remember the behavior of a user and predict the next service required by the user successfully. In this research, a recently developed algorithm named as sequence prediction via enhanced episode discovery (SPEED) is considered for modification by inclusion of location agents. A smart home prototype consisting of two rooms is designed as a testbed for verification. The results show that the accuracy of this algorithm is more than 40%, which is better than the previous SPEED. Moreover, the algorithm detects the location of next predicted event. Since human activity can be distinguished by their existing locations, predicting the next event as well as the location helps to determine the next action more accurately.

AB - Smart home or home automation has become widely popular especially in the case of easing the lives of people with special needs, for instance the elderly and handicapped people. In every home, a specific user has a unique pattern or sequence of using the functions of that house. Recognizing that unique pattern is the key to ensuring an intelligently and properly automated household where the house will remember the behavior of a user and predict the next service required by the user successfully. In this research, a recently developed algorithm named as sequence prediction via enhanced episode discovery (SPEED) is considered for modification by inclusion of location agents. A smart home prototype consisting of two rooms is designed as a testbed for verification. The results show that the accuracy of this algorithm is more than 40%, which is better than the previous SPEED. Moreover, the algorithm detects the location of next predicted event. Since human activity can be distinguished by their existing locations, predicting the next event as well as the location helps to determine the next action more accurately.

KW - Artificial intelligence

KW - Hidden markov models

KW - Location agent

KW - Prediction by partial matching

KW - Smart homes

UR - http://www.scopus.com/inward/record.url?scp=85020283167&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020283167&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:85020283167

VL - 10

SP - 161

EP - 165

JO - Journal of Engineering Science and Technology Review

JF - Journal of Engineering Science and Technology Review

SN - 1791-9320

IS - 2

ER -