1. UKM Learning and Research Repository
  2. Master Thesis / Tesis Master
  3. Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina
Please use this identifier to cite or link to this item: https://ptsldigital.ukm.my/jspui/handle/123456789/457847
Full metadata record
DC FieldValueLanguage
dc.contributor.advisorMd Mamun Ibne Reaz, Prof. Dr.
dc.contributor.authorAraf Farayez (P90055)
dc.date.accessioned2023-09-12T09:14:16Z-
dc.date.available2023-09-12T09:14:16Z-
dc.date.issued2020-01-23
dc.identifier.otherukmvital:123623
dc.identifier.urihttps://ptsldigital.ukm.my/jspui/handle/123456789/457847-
dc.descriptionSmart home systems have become major contributors in automation, healthcare, and security. An important role of these systems is to record user behaviors and automate daily activities that can be particularly useful in elder care and homes for disabled people. Activity prediction algorithms are used extensively in smart homes for generating resident behaviour models. These algorithms use machine learning algorithms to generate data models from resident actions and use these models to predict next actions. But recent activity prediction approaches need further research before being implemented in real life scenarios. This thesis focuses on improving the prediction accuracy of smart home systems by introducing a novel activity learning method based on a popular data compression technique. The proposed algorithm, Activity Prediction using Partial Matching uses a prefix tree-based data model in order to learn and predict user actions. The algorithm applies episode discovery to detect correlated sensor events and learns the activities using a lossless data compression technique called Prediction by Partial Matching. This process assigns a probability of occurrence to sensor events and uses these probabilities to detect patterns in resident behaviour. The results are compared with several recent activity prediction algorithms to determine the efficiency in terms of accuracy, memory usage, and runtime. In terms of accuracy, 8.22% improvement is achieved from its predecessors. The algorithm has been tested on large datasets collected by renowned smart home test beds. It achieves 66.69% better memory efficiency and 37.00% faster runtime on large datasets. This algorithm can be effectively implemented in smart homes to improve the activity prediction and deliver reliable automation.,Master of Science
dc.language.isoeng
dc.publisherUKM, Bangi
dc.relationFaculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina
dc.rightsUKM
dc.subjectUniversiti Kebangsaan Malaysia -- Dissertations
dc.subjectDissertations, Academic -- Malaysia
dc.subjectAlgorithm
dc.subjectSmart home
dc.subjectAutomation
dc.titleActivity prediction algorithm using prefix tree - based context generation for smart home automation
dc.typetheses
dc.format.pages143
dc.identifier.barcode005797(2021)(PL2)
Appears in Collections:Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina

Files in This Item:
File Description SizeFormat 
ukmvital_123623+SOURCE1+SOURCE1.0.PDF
  Restricted Access		2.85 MBAdobe PDFThumbnail
View/Open
Show simple item record Recommend this item


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.