Siti Nur Hidayah, Mazelan (2022) Abnormal Pattern Detection In Ppg Signals Using Time Series Analysis. College of Engineering, Universiti Malaysia Pahang Al-Sultan Abdullah.
|
Pdf
EA18084_Hidayah_ThesisV2 - Hidayah Mazelan.pdf - Accepted Version Download (1MB) | Preview |
Abstract
The photoplethysmogram (PPG) signal is a data in continuous real-time series. It depicts the peripheral pulse wave that is produced due to heart activity, respiration, and other physiological effects. The time-series signal contains a lot of information which is difficult to be processed. The abnormal PPG signal is messy, non-periodic, and irregular. Several existing methods such as Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Deep Neural Network (DNN) and sensor had been used to detect abnormal pattern from PPG signal which can produce high performance and accuracy. However, these methods are higher in complexity or have uncertain repeatability. Therefore, this thesis proposed a method which is rule-based algorithm that is less complex, with quicker and more simple training, reducing the errors while still producing high accuracy. This project’s objectives are to implement rule-based algorithm method for abnormal pattern detection in PPG signals, and to investigate the accuracy and performance of rule-based algorithm in detecting the abnormal pattern. The signal processing, segmentation, feature extraction, training and testing for rule-based algorithm classifier, using wrist PPG during exercise dataset and pulse transmit time dataset, are done in this study to detect the abnormal pattern in PPG signals. The accuracy and coverage of rule for both training and testing process are recorded in order to determine the performance of the method used in this study. The abnormal PPG pattern detection using rule-based algorithm has produced accuracy of 87.30% in training process and 87.18% in testing process with coverage of rule for training and testing, 89.26% and 87.33%. The findings of this project can be further used for application of abnormal pattern in PPG signal such as healthcare and human activity recognition.
| Item Type: | Undergraduates Project Papers |
|---|---|
| Additional Information: | SV: Ts. Dr. Asrul bin Adam |
| Uncontrolled Keywords: | photoplethysmogram (PPG), Convolutional Neural Network (CNN), Recurrent Neural Network (RNN), Deep Neural Network (DNN) |
| Subjects: | T Technology > TA Engineering (General). Civil engineering (General) T Technology > TK Electrical engineering. Electronics Nuclear engineering |
| Faculty/Division: | College of Engineering |
| Depositing User: | Mr. Nik Ahmad Nasyrun Nik Abd Malik |
| Date Deposited: | 08 Jan 2024 06:36 |
| Last Modified: | 08 Jan 2024 06:36 |
| URI: | http://umpir.ump.edu.my/id/eprint/39894 |
| Download Statistic: | View Download Statistics |
Actions (login required)
 |
View Item |