Conditional Random Field Feature Generation of Smart Home Sensor Data using Random Forests

Mark Eastwood, Alexandros Konios, Bo Tan, Yanguo Jing, Abdul Hamid

Research output: Chapter in Book/Report/Conference proceedingConference proceeding

Abstract

A typical approach to building a feature set for a conditional
random field model is to build a large set of conjunctions of atomic tests, all of which adhere to a small number of relatively simple templates. Building more complex features in this way can be difficult, as the more complex templates needed to do this can result in a combinatoric explosion in the number of features. We use the inherent instability of decision trees to produce a small set of more complex conjunctions that are particularly suitable for the problem to be solved, using the same techniques used in generating random forest ensemble classifiers, and
build a CRF on these features. We apply this method to an activity
recognition problem on a dataset from the CASAS smart home project, in which we predict activities of daily living from sensor activations.
LanguageEnglish
Title of host publication2019 International Microwave Biomedical Conference (IMBioC 2019)
PublisherIEEE
Pages(In-press)
Volume(In-press)
Publication statusAccepted/In press - 26 Jan 2019
EventIEEE International Microwave Biomedical Conference - China, Nanjing, China
Duration: 6 May 20198 May 2019
http://www.em-conf.com/imbioc2019/conference/html.php?title=Registration

Conference

ConferenceIEEE International Microwave Biomedical Conference
CountryChina
CityNanjing
Period6/05/198/05/19
Internet address

Fingerprint

Decision trees
Explosions
Classifiers
Chemical activation
Sensors

Keywords

  • Smart home
  • machine learning
  • assistant living

Cite this

Eastwood, M., Konios, A., Tan, B., Jing, Y., & Hamid, A. (Accepted/In press). Conditional Random Field Feature Generation of Smart Home Sensor Data using Random Forests. In 2019 International Microwave Biomedical Conference (IMBioC 2019) (Vol. (In-press), pp. (In-press)). IEEE.

Conditional Random Field Feature Generation of Smart Home Sensor Data using Random Forests. / Eastwood, Mark; Konios, Alexandros; Tan, Bo; Jing, Yanguo; Hamid, Abdul.

2019 International Microwave Biomedical Conference (IMBioC 2019). Vol. (In-press) IEEE, 2019. p. (In-press).

Research output: Chapter in Book/Report/Conference proceedingConference proceeding

Eastwood, M, Konios, A, Tan, B, Jing, Y & Hamid, A 2019, Conditional Random Field Feature Generation of Smart Home Sensor Data using Random Forests. in 2019 International Microwave Biomedical Conference (IMBioC 2019). vol. (In-press), IEEE, pp. (In-press), IEEE International Microwave Biomedical Conference, Nanjing, China, 6/05/19.
Eastwood M, Konios A, Tan B, Jing Y, Hamid A. Conditional Random Field Feature Generation of Smart Home Sensor Data using Random Forests. In 2019 International Microwave Biomedical Conference (IMBioC 2019). Vol. (In-press). IEEE. 2019. p. (In-press)
Eastwood, Mark ; Konios, Alexandros ; Tan, Bo ; Jing, Yanguo ; Hamid, Abdul. / Conditional Random Field Feature Generation of Smart Home Sensor Data using Random Forests. 2019 International Microwave Biomedical Conference (IMBioC 2019). Vol. (In-press) IEEE, 2019. pp. (In-press)
@inproceedings{32fe233bd3414f0db7080bcdd0d34333,
title = "Conditional Random Field Feature Generation of Smart Home Sensor Data using Random Forests",
abstract = "A typical approach to building a feature set for a conditionalrandom field model is to build a large set of conjunctions of atomic tests, all of which adhere to a small number of relatively simple templates. Building more complex features in this way can be difficult, as the more complex templates needed to do this can result in a combinatoric explosion in the number of features. We use the inherent instability of decision trees to produce a small set of more complex conjunctions that are particularly suitable for the problem to be solved, using the same techniques used in generating random forest ensemble classifiers, andbuild a CRF on these features. We apply this method to an activityrecognition problem on a dataset from the CASAS smart home project, in which we predict activities of daily living from sensor activations.",
keywords = "Smart home, machine learning, assistant living",
author = "Mark Eastwood and Alexandros Konios and Bo Tan and Yanguo Jing and Abdul Hamid",
year = "2019",
month = "1",
day = "26",
language = "English",
volume = "(In-press)",
pages = "(In--press)",
booktitle = "2019 International Microwave Biomedical Conference (IMBioC 2019)",
publisher = "IEEE",

}

TY - GEN

T1 - Conditional Random Field Feature Generation of Smart Home Sensor Data using Random Forests

AU - Eastwood, Mark

AU - Konios, Alexandros

AU - Tan, Bo

AU - Jing, Yanguo

AU - Hamid, Abdul

PY - 2019/1/26

Y1 - 2019/1/26

N2 - A typical approach to building a feature set for a conditionalrandom field model is to build a large set of conjunctions of atomic tests, all of which adhere to a small number of relatively simple templates. Building more complex features in this way can be difficult, as the more complex templates needed to do this can result in a combinatoric explosion in the number of features. We use the inherent instability of decision trees to produce a small set of more complex conjunctions that are particularly suitable for the problem to be solved, using the same techniques used in generating random forest ensemble classifiers, andbuild a CRF on these features. We apply this method to an activityrecognition problem on a dataset from the CASAS smart home project, in which we predict activities of daily living from sensor activations.

AB - A typical approach to building a feature set for a conditionalrandom field model is to build a large set of conjunctions of atomic tests, all of which adhere to a small number of relatively simple templates. Building more complex features in this way can be difficult, as the more complex templates needed to do this can result in a combinatoric explosion in the number of features. We use the inherent instability of decision trees to produce a small set of more complex conjunctions that are particularly suitable for the problem to be solved, using the same techniques used in generating random forest ensemble classifiers, andbuild a CRF on these features. We apply this method to an activityrecognition problem on a dataset from the CASAS smart home project, in which we predict activities of daily living from sensor activations.

KW - Smart home

KW - machine learning

KW - assistant living

M3 - Conference proceeding

VL - (In-press)

SP - (In-press)

BT - 2019 International Microwave Biomedical Conference (IMBioC 2019)

PB - IEEE

ER -