Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/497729Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Saadiyah Darus, Assoc. Prof. Dr. | - |
| dc.contributor.author | Norwati Md Yusof (P48370) | - |
| dc.date.accessioned | 2023-10-13T08:08:48Z | - |
| dc.date.available | 2023-10-13T08:08:48Z | - |
| dc.date.issued | 2019-12-09 | - |
| dc.identifier.other | ukmvital:117657 | - |
| dc.identifier.uri | https://ptsldigital.ukm.my/jspui/handle/123456789/497729 | - |
| dc.description | The challenge of reconciling translation studies with machine translation now has a new paradigm where human evaluation has been considered in automated translation. However, statistical approaches dominate machine translation while models in translation studies generally describe macro approaches focusing on culture and linguistic elements. To reconcile translation studies with machine translation, both areas need to consider translation criteria proper and to involve human evaluators. This study aims to develop a micro level model based on human perceptions for English to Malay translation evaluation and to test it with the performance of an English to Malay machine translation system (SISTEC). The research scope focuses on examining the evaluation criteria for two attributes, namely Accuracy and Intelligibility. The study uses a mixed method approach with soft data and hard data. It begins with researching and identifying the attributes and the criteria for evaluation through literature review. However, only three criteria for each attribute were further selected through results from interviews and incorporated as part of a questionnaire to find the range of criteria for evaluation according to human evaluators. The questionnaire was also designed to evaluate the performance of SISTEC by using randomly selected sentences from the system output which were previously extracted from an established corpus. Results of the study are significant in two ways. First, the expectation of human evaluators for Accuracy and Intelligibility in both human and machine translation is high. Based on Likert Scale of 1 to 4, the result shows that all the criteria are ranged between the scale of 3 to 4. The expected range for all the criteria for machine translation never crossed or reached higher than the range expected for human translation. The findings also show that SISTEC achieved only 20 percent of the mean for all the criteria. The study has two significant implications; in translation studies and machine translation, it contributes a micro level evaluation model with the established criteria based on human evaluators, while in automated translation for English to Malay, the study observed that the performance of SISTEC can be upgraded in terms of Accuracy and Intelligibility.,'Certification of Masters/Doctoral Thesis' is not available,Ph.D. | - |
| dc.language.iso | eng | - |
| dc.publisher | UKM, Bangi | - |
| dc.relation | Faculty of Social Sciences and Humanities / Fakulti Sains Sosial dan Kemanusiaan | - |
| dc.rights | UKM | - |
| dc.subject | Machine translating | - |
| dc.subject | Computational linguistics | - |
| dc.subject | Universiti Kebangsaan Malaysia -- Dissertations | - |
| dc.subject | Dissertations, Academic -- Malaysia | - |
| dc.title | Malaysian evaluators' perspective on evaluation for English to Malay machine translation | - |
| dc.type | Theses | - |
| dc.format.pages | 334 | - |
| dc.identifier.callno | P308.N637 2019 tesis | - |
| dc.identifier.barcode | 004508(2020) | - |
| Appears in Collections: | Faculty of Social Sciences and Humanities / Fakulti Sains Sosial dan Kemanusiaan | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ukmvital_117657+SOURCE1+SOURCE1.0.PDF Restricted Access | 1.04 MB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.