Hybrid neural network in medicolegal degree of injury determination based on Visum et Repertum
Essentially, determination model for degree of injury is crucial for refining diagnostic and increasing accuracy of the Forensic and Medicolegal services. Existing models are deemed difficult in identifying the critical features. These are due to the model having insufficient of critical features an...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | en |
| Published: |
2023
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/28268/2/Hybrid%20neural%20network%20in%20medicolegal%20degree%20of%20injury%20determination%20based%20on%20Visum%20et%20Repertum.pdf http://eprints.utem.edu.my/id/eprint/28268/ https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=123811 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1859054981180555264 |
|---|---|
| author | Wardhana, Mohammad Hadyan |
| author_facet | Wardhana, Mohammad Hadyan |
| author_sort | Wardhana, Mohammad Hadyan |
| building | UTEM Library |
| collection | Institutional Repository |
| content_provider | Universiti Teknikal Malaysia Melaka |
| content_source | UTEM Institutional Repository |
| continent | Asia |
| country | Malaysia |
| description | Essentially, determination model for degree of injury is crucial for refining diagnostic and increasing accuracy of the Forensic and Medicolegal services. Existing models are deemed difficult in identifying the critical features. These are due to the model having insufficient of critical features analysis that cause the inconsistency decision to determine degree of injury among the medical practitioners. The issue become more complex because the dataset consists of incomplete data and outliers class problem that can affects the sampling bias. The purpose of this study is to identify the characteristics and terms, develop and evaluate the Hybrid Neural Network Model (HNNM) for determining degree of injury based on Visum et Repertum (VeR) data. The VeR data consist of 289 patients’ record. The HNNM is expected to determine either the persecution victim having a minor, moderate, or serious injury which inclusively mention in Indonesian Penal Code. HNNM is developed based on the case studies at three hospitals in Pekanbaru comprise three main phases which are pre-processing, development, and performance analysis. Pre-processing phase overcomes the issue of incomplete data by performing data cleansing and data normalization. The development phase begins with utilizing Analytical Hierarchical Process (AHP) to validate the ranking for each of weight on the critical features from the experts’ opinion. Then, the selection of the critical features is chosen via Neural Network (NN) as classification algorithm and Genetic Algorithm (GA) as an optimization technique. The selected critical features are applied during the dataset training stages to improve the accuracy and reduce error of the HNNM. GA is aimed to increase the accuracy and minimize the error in the learning stages of NN. The development phase accomplished with testing stages by employing VeR dataset. The performance analysis shows the HNNM produced 98.85% accuracy level and Root Mean Square Error (RMSE) value at 0.077. In the validation stage, the questionnaires are answered by the Subject Matter Expert (SME) groups which consist of feature, implementation, and viability aspect of HNNM. Result from the questionnaires concluded that the agreement level of SMEs reaches up to 80%. Thus, the features of the HNNM are implementable and highly acceptable by the practitioner. For the future research, the HNNM need to increase the accuracy by improving the input features including lifestyle, habit, and job. |
| format | Thesis |
| id | my.utem.eprints-28268 |
| institution | Universiti Teknikal Malaysia Melaka |
| language | en |
| publishDate | 2023 |
| record_format | eprints |
| spelling | my.utem.eprints-282682026-03-05T01:08:51Z http://eprints.utem.edu.my/id/eprint/28268/ Hybrid neural network in medicolegal degree of injury determination based on Visum et Repertum Wardhana, Mohammad Hadyan Q Science (General) QA Mathematics Essentially, determination model for degree of injury is crucial for refining diagnostic and increasing accuracy of the Forensic and Medicolegal services. Existing models are deemed difficult in identifying the critical features. These are due to the model having insufficient of critical features analysis that cause the inconsistency decision to determine degree of injury among the medical practitioners. The issue become more complex because the dataset consists of incomplete data and outliers class problem that can affects the sampling bias. The purpose of this study is to identify the characteristics and terms, develop and evaluate the Hybrid Neural Network Model (HNNM) for determining degree of injury based on Visum et Repertum (VeR) data. The VeR data consist of 289 patients’ record. The HNNM is expected to determine either the persecution victim having a minor, moderate, or serious injury which inclusively mention in Indonesian Penal Code. HNNM is developed based on the case studies at three hospitals in Pekanbaru comprise three main phases which are pre-processing, development, and performance analysis. Pre-processing phase overcomes the issue of incomplete data by performing data cleansing and data normalization. The development phase begins with utilizing Analytical Hierarchical Process (AHP) to validate the ranking for each of weight on the critical features from the experts’ opinion. Then, the selection of the critical features is chosen via Neural Network (NN) as classification algorithm and Genetic Algorithm (GA) as an optimization technique. The selected critical features are applied during the dataset training stages to improve the accuracy and reduce error of the HNNM. GA is aimed to increase the accuracy and minimize the error in the learning stages of NN. The development phase accomplished with testing stages by employing VeR dataset. The performance analysis shows the HNNM produced 98.85% accuracy level and Root Mean Square Error (RMSE) value at 0.077. In the validation stage, the questionnaires are answered by the Subject Matter Expert (SME) groups which consist of feature, implementation, and viability aspect of HNNM. Result from the questionnaires concluded that the agreement level of SMEs reaches up to 80%. Thus, the features of the HNNM are implementable and highly acceptable by the practitioner. For the future research, the HNNM need to increase the accuracy by improving the input features including lifestyle, habit, and job. 2023 Thesis NonPeerReviewed text en cc_by_4 http://eprints.utem.edu.my/id/eprint/28268/2/Hybrid%20neural%20network%20in%20medicolegal%20degree%20of%20injury%20determination%20based%20on%20Visum%20et%20Repertum.pdf Wardhana, Mohammad Hadyan (2023) Hybrid neural network in medicolegal degree of injury determination based on Visum et Repertum. Doctoral thesis, Universiti Teknikal Malaysia Melaka. https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=123811 |
| spellingShingle | Q Science (General) QA Mathematics Wardhana, Mohammad Hadyan Hybrid neural network in medicolegal degree of injury determination based on Visum et Repertum |
| title | Hybrid neural network in medicolegal degree of injury determination based on Visum et Repertum |
| title_full | Hybrid neural network in medicolegal degree of injury determination based on Visum et Repertum |
| title_fullStr | Hybrid neural network in medicolegal degree of injury determination based on Visum et Repertum |
| title_full_unstemmed | Hybrid neural network in medicolegal degree of injury determination based on Visum et Repertum |
| title_short | Hybrid neural network in medicolegal degree of injury determination based on Visum et Repertum |
| title_sort | hybrid neural network in medicolegal degree of injury determination based on visum et repertum |
| topic | Q Science (General) QA Mathematics |
| url | http://eprints.utem.edu.my/id/eprint/28268/2/Hybrid%20neural%20network%20in%20medicolegal%20degree%20of%20injury%20determination%20based%20on%20Visum%20et%20Repertum.pdf http://eprints.utem.edu.my/id/eprint/28268/ https://plh.utem.edu.my/cgi-bin/koha/opac-detail.pl?biblionumber=123811 |
| url_provider | http://eprints.utem.edu.my/ |