Timor-Leste Journal of Engineering and Science

Classification of Tetun Language Documents Based on INL and DIT Orthography with a Text Mining Approach

2024-08-08T03:22:14+00:00

The main problem in language classification is the complexity and intricacy of accurately tracing these relationships such as language evolution, contact and borrowing words which makes it difficult to classify the orthography used. In both government and non-government institutions in the country, many individuals write documents using varying spellings. Currently, at the Dili Institute of Technology (DIT), a unique spelling system has been developed alongside adherence to the guidelines of the National Institute of Linguistics (INL). The DIT orthography, which is based on contemporary Tetun, does not employ accents, as numerous studies have indicated that accents are unnecessary. The objective of this research is to develop an application that classifies documents using a text mining approach, with tokenization and filtering based on word lists from INL and DIT orthographies. This process aims to accurately categorize submitted user documents. The documents used in this research consists of INL and DIT orthographic. The word list dictionary from INL comprises 1,487 words from the Tetun-Portuguese dictionary, while the DIT word list includes 756 words collected from the DIT Language Center and additional sources. The research findings indicate that the system is capable of classifying documents based on the predefined orthographic categories.

Interpreting Well And Reservoir Model Based On The Well Testing Approach For Constructing Well Productivity

2024-07-24T23:51:06+00:00

Well testing is key to constructing the reservoir model, especially in the development of fields. Well test well-known as pressure transient analysis seeks the dynamic behavior of a reservoir in an inverse problem manner. The pressure transient analysis measures the change in pressure at the wellbore by altering the production rate that can provide a signature of reservoir properties typically in the build-up period. Availability of data is run into the first simulator Ecrin V4 thoroughly monitoring the change in pressure data to the production rate. Pressure and its derivative which is derived from the diffusivity equation are compared to reveal both models in a system of well production. Results show that the skin has negativity -3.43 to refer as no damage and 0.0125 bbl/d of wellbore coefficient at the vicinity wellbore. Further, dual porosity is identified as the reservoir model in which the derivative response showed the transitional dip at the middle time, and aside from that the infinite boundary act flattened at late time. To conclude, the initial pressure of 3915.35 psi in the matrix block flows into the fissure system with an average permeability of 100.8 md. An average pressure in the fissure system can be estimated using the transient flow equation which suits pressure drop depending on the radius and time. Once the reservoir pressure is estimated, 3900 psi. It is necessary to construct the well productivity. The second simulator Pipesim is used to design the inflow performance relationship and the tubing performance. The IPR was continued with Vogel to consider gas dissolved of 400 scf/stb and the tubing was assumed with an inside diameter of 2.735. Finally, the well production may be known as about 32% of AOF 18505.7 stb/d. This interpretation is simple and applicable to unlocking the well and reservoir model for constructing the well productivity-based computational model.

Optimization of Early Warning System for Landslides Based on Rainfall Using Naive Bayes Classifier and Multiclass Support Vector Machine Algorithm in Takari Region

2024-07-24T23:51:07+00:00

This study explores non-structural disaster mitigation approaches employed by researchers, utilizing machine learning algorithms to analyze weather data and assess landslide vulnerability in the Takari Sub-district. Through field investigations and secondary data analysis, the research underscores the significance of rainfall intensity as a key factor in triggering landslides in the region. Additionally, soil types and slope gradients are identified as critical considerations in landslide vulnerability detection. Evaluation of a multiclass support vector algorithm for rainfall prediction reveals a notable accuracy rate of 57.97%, with predictions indicating instances of various rainfall intensities. Factors influencing these predictions include average temperature, humidity, wind speed, duration of sunshine, and wind direction. However, the study notes limitations in predictive accuracy due to the constrained availability of rainfall data. Consequently, the findings emphasize the need for preemptive measures, urging governmental authorities and local communities to prioritize structural disaster mitigation strategies to mitigate the heightened susceptibility to landslides in the Takari region

Analytic Hierarchy Process and Multi-Factor Evaluation Process Methods for Proposal Research Evaluation

2024-07-24T23:51:07+00:00

Evaluation of scientific research proposals is one of the most important factors in determining the quality of research results to be obtained so that they can contribute to the development of science and technology in harmony with people's lives, so this needs to be considered. However, the process of evaluating scientific research proposals is not an easy matter because it involves a variety of complex causal factors and sub-factors in conducting evaluations in a consistent and objective manner. Therefore, we propose the AHP and MFEP methods with causal factors, namely originality, novelty, contribution, methodology, reputation journal references, research roadmap, research member, up-to-date references, percentage of references, tools references, references styles, proposal format, and thirty-three (33) other sub-factors. This study aims to provide knowledge about how the AHP and MFEP methods can be combined to evaluate scientific research Proposals. The AHP method is used to calculate the weight of the priority level values for each causal factor and sub-factor that will be used by the MFEP method, while the MFEP method is used to calculate the evaluation weight value for each alternative by utilizing the value of the priority level of causal factors and sub-determining factors resulting from AHP, as well as calculating the total value of the evaluation weight for each alternative. The results showed that both methods can be used to evaluate scientific research proposals by obtaining five (5) alternative candidates for research grant winners from CARPS-CS with a total evaluation weight value of = 50% out of ten (10) alternative candidates.

ChatGPT in Education: A Comprehensive Examination of its Impact on Student Learning and Achievement

2024-07-24T23:51:07+00:00

The adoption of artificial intelligence (AI) technologies like ChatGPT in educational settings has sparked significant interest and discussion. ChatGPT offers personalized feedback and explanations customized to individual student needs, with the goal of enhancing engagement, comprehension, and academic performance. The objective of research to review engages with ongoing debates surrounding AI in education, particularly concerning issues of critical thinking, creativity, and the evolving role of educators in AI-driven learning environments. The review method employed in this research was synthesize existing research involved conducting a systematic search of academic databases using relevant keywords ("ChatGPT," "AI in education," "student achievement") to identify peer-reviewed articles, conference papers, and reports. Result of synthesize shows the integration of AI-driven tools like ChatGPT into educational settings has the potential to significantly enhance student achievement by tailoring learning experiences, refining language skills, cultivating critical thinking, providing supplementary resources, offering timely feedback, and boosting motivation and confidence.