Continual learning and data analysis of time series data
Time Series Data (TSD) has become the cornerstone of critical applications in various fields. However, temporal analysis faces a significant challenge called catastrophic forgetting, where previously acquired knowledge or skills are lost when learning new tasks. Therefore, this study aims to integra...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/176923 |
| _version_ | 1811690257189437440 |
|---|---|
| author | Ke, Tangxin |
| author2 | Soh Yeng Chai |
| author_facet | Soh Yeng Chai Ke, Tangxin |
| author_sort | Ke, Tangxin |
| collection | NTU |
| description | Time Series Data (TSD) has become the cornerstone of critical applications in various fields. However, temporal analysis faces a significant challenge called catastrophic forgetting, where previously acquired knowledge or skills are lost when learning new tasks. Therefore, this study aims to integrate continuous learning (CL) techniques to mitigate the phenomenon of catastrophic forgetting and enhance the model's capacity for processing TSD. Focusing on the Human Activity Recognition (HAR) problem, this study utilized a hybrid CNN-LSTM hybrid model as the baseline and explored a range of continuous learning techniques, including Experience Replay, Elastic Weight Consolidation (EWC) and Progressive Neural Network (PNN) etc. to show the effectiveness of CL. |
| first_indexed | 2024-10-01T06:01:07Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/176923 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T06:01:07Z |
| publishDate | 2024 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1769232024-05-24T15:44:02Z Continual learning and data analysis of time series data Ke, Tangxin Soh Yeng Chai School of Electrical and Electronic Engineering EYCSOH@ntu.edu.sg Engineering Time Series Data (TSD) has become the cornerstone of critical applications in various fields. However, temporal analysis faces a significant challenge called catastrophic forgetting, where previously acquired knowledge or skills are lost when learning new tasks. Therefore, this study aims to integrate continuous learning (CL) techniques to mitigate the phenomenon of catastrophic forgetting and enhance the model's capacity for processing TSD. Focusing on the Human Activity Recognition (HAR) problem, this study utilized a hybrid CNN-LSTM hybrid model as the baseline and explored a range of continuous learning techniques, including Experience Replay, Elastic Weight Consolidation (EWC) and Progressive Neural Network (PNN) etc. to show the effectiveness of CL. Bachelor's degree 2024-05-23T06:41:21Z 2024-05-23T06:41:21Z 2024 Final Year Project (FYP) Ke, T. (2024). Continual learning and data analysis of time series data. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/176923 https://hdl.handle.net/10356/176923 en B1096-231 application/pdf Nanyang Technological University |
| spellingShingle | Engineering Ke, Tangxin Continual learning and data analysis of time series data |
| title | Continual learning and data analysis of time series data |
| title_full | Continual learning and data analysis of time series data |
| title_fullStr | Continual learning and data analysis of time series data |
| title_full_unstemmed | Continual learning and data analysis of time series data |
| title_short | Continual learning and data analysis of time series data |
| title_sort | continual learning and data analysis of time series data |
| topic | Engineering |
| url | https://hdl.handle.net/10356/176923 |
| work_keys_str_mv | AT ketangxin continuallearninganddataanalysisoftimeseriesdata |