USU-IR Repositori Institusi Universitas Sumatera Utara
Not a member yet
100433 research outputs found
Sort by
The Influence of Organization Citizenship Behavior and Employee Psychological Capital on Employee Performance at Bank Sumut With job satisfaction as a moderation variable
No full textIn an increasingly dynamic work environment, employees' ability to voluntarily contribute through Organizational Citizenship Behavior (OCB) and possess high Psychological Capital, such as optimism, hope, resilience, and self-efficacy, is considered a critical factor influencing performance. Moreover, job satisfaction is often regarded as a key element that strengthens the relationship between these variables and performance. This study aims to analyze the influence of Organizational Citizenship Behavior (OCB) and Psychological Capital on employee performance at Bank Sumut, with job satisfaction as a moderating variable. A quantitative approach was employed in this study. The sample consisted of 83 employees out of 473 employees at Bank Sumut, selected using a simple random sampling technique. Data analysis was conducted using Structural Equation Modeling (SEM). The results of this study indicate that Organizational Citizenship Behavior (OCB) positively influences employee performance, Psychological Capital positively influences employee performance, job satisfaction weakens the effect of Organizational Citizenship Behavior on employee performance, and job satisfaction strengthens the effect of Psychological Capital on the performance of Bank Sumut employees.192 PagesTesis Magiste
Deepfake Detection on Faces in Video with Cross-Attention Multi-Scale Vision Transformer and EfficientNet
No full textThe proliferation of sophisticated deepfake videos poses a serious threat to digital trust and security, demanding detection systems that are not only accurate but also computationally efficient for practical applications. This research aims to design, implement, and evaluate a hybrid architecture that balances high accuracy with inference efficiency for video-based deepfake detection. The proposed model integrates EfficientNet-B1 as a feature extractor with a Cross-Attention Multi-Scale Vision Transformer (Cross-ViT) for context modeling. The model was trained on a combination of the FaceForensics++ and Celeb-DF(v2) datasets and evaluated on an out-of-distribution dataset, DeepFakeDetection (DFD), to test its generalization capabilities. The evaluation results demonstrate reliable detection performance, achieving an Area Under the Curve (AUC) of 92.35% and a video-level F1-score of 83.62%. The model's primary advantage is its exceptional computational efficiency, requiring only 0.349 G-FLOP for per-frame inference, despite having a large parameter count (114.33 Million). This study also reveals that the use of a small batch size, Face-Cutout augmentation, and a Binary Cross-Entropy (BCE) loss function significantly contributes to improved generalization and effective video-level aggregation. This research successfully validates an efficient and scalable hybrid architecture that offers a practical solution for deepfake detection by balancing accuracy, inference speed, and model size.85 PagesSkripsi Sarjan
Decentralization of Education through Local Content Curriculum in Preserving Local Wisdom at the Education Office of North Sumatra Province
No full textThe decentralization of education grants the North Sumatra Provincial Education Office the authority to develop local content curriculum in accordance with local potential and needs, with the aim of preserving local wisdom amidst the tides of modernization and globalization. However, in its implementation, there are still issues such as limited resources, low stakeholder participation, and the lack of readiness among schools and teachers. This research aims to provide an in-depth description of the decentralization of education through local content curriculum in preserving local wisdom at the Education Office of North Sumatra Province. The research method used is descriptive qualitative. The research informants include the Head of the Curriculum Section and Education Quality Analyst in the Senior High School Development Division of the North Sumatra Provincial Education Office, the Deputy Principal for Curriculum and teachers, as well as students and their parents. The data and information that have been obtained will be analyzed using Lindblom's incremental model with three indicators, namely adjustment to the conditions and needs of the community, continuous renewal, and participation of actors and stakeholders. Research results indicate that normatively, provincial governments have the authority to implement local content curriculum. However, in reality, it has not been implemented as it should be. This can be seen in high schools that do not implement local content lessons because they are unaware of the governor's decision. In the end, the school implemented the local content curriculum through the Pancasila Student Profile Strengthening Project (P5), which is a central policy. On the indicator of adjustment to the conditions and needs of the community, the availability of classrooms and teaching materials is adequate, but there are no special local content teachers due to budget constraints. Next, continuous renewal provides room for innovation, but in practice, it still relies on the central government. Lastly, community participation is still limited. The suggestion given is the need for openness of information regarding local content curriculum policies to schools, provision of adequate resources, and community involvement so that the decentralization of education can preserve the local wisdom in North Sumatra.201 PagesSkripsi Sarjan
PM2.5 and SO₂ Monitoring System and Acid Rain Impact Analysis on Soil pH around Pangkalan Susu Power Plant Integrated with Internet of Things (IoT)”
No full textAir pollution caused by sulfur dioxide (SO₂) emissions and fine particulate matter
(PM2.5) had the potential to cause acid rain, which affected soil quality, particularly
soil pH. This study aimed to design and develop a microcontroller-based air quality
monitoring system integrated with Internet of Things (IoT) technology to monitor
PM2.5 and SO₂ concentrations as well as soil pH in real-time around the Pangkalan
Susu coal-fired power plant (PLTU). The system utilized GP2Y1010AU0F, MQ-136,
and soil pH sensors, with data visualized through the Blynk platform. A 14-day
monitoring period was conducted, and the results showed that PM2.5 and SO₂ levels
in Tanjung Pasir were significantly higher than in Lubuk Kasih, with PM2.5 peaking
at 176.55 µg/m³ and SO₂ at 4 ppm. This resulted in a decrease in soil pH in Tanjung
Pasir to an average of < 5.8, while Lubuk Kasih remained stable above 6.5. Pearson
correlation tests indicated a strong and significant negative relationship between
PM2.5, SO₂, and soil pH in Tanjung Pasir (r = –0.750 and –0.798; p < 0.01). This
study demonstrated that the designed system was capable of detecting acid rain
potential and its impact on soil, and showed potential for long-term environmental
monitoring applications.175 PagesSkripsi Sarjan
Analysis Of Mobile Banking User Activity Based On Transaction Time Clustering Using Self-Organizing MAP (SOM) Method
No full textPertumbuhan pesat layanan mobile banking di Indonesia menuntut pemahaman yang
lebih dalam terhadap perilaku pengguna, terutama dalam hal waktu dan pola transaksi.
Namun, tantangan yang dihadapi adalah bagaimana mengelompokkan pengguna secara
efektif berdasarkan kebiasaan waktu mereka dalam melakukan transaksi, agar strategi
layanan dapat disesuaikan secara tepat sasaran. Untuk menjawab permasalahan
tersebut, penelitian ini menerapkan metode Self-Organizing Maps (SOM) guna
melakukan klasterisasi terhadap pengguna berdasarkan fitur waktu transaksi, seperti
jumlah transaksi pagi, siang, sore, malam, serta pembagian antara hari kerja dan akhir
pekan. Dataset yang digunakan mencakup 87.361 pengguna mobile banking sepanjang
tahun 2023. Hasil penelitian menunjukkan bahwa metode SOM mampu membentuk
sembilan klaster perilaku pengguna yang berbeda, dengan klaster terbesar adalah Early
User (Weekday) yang terdiri dari 32.324 pengguna (37,0%). Secara keseluruhan,
segmen Early User (Weekday) mencakup sekitar 60,3% populasi pengguna. Sementara
itu, terdapat juga segmen minoritas seperti Night Owl (Weekday) (5,9%) dan Early User
(Weekend) (2,7%) yang menunjukkan pola perilaku unik. Evaluasi performa model
menghasilkan nilai Quantization Error (QE) sebesar 0,339 dan Topographic Error (TE)
sebesar 0,066, baik pada data validasi maupun data pengujian, menandakan bahwa
hasil klasterisasi cukup akurat dan topologi pemetaan data terjaga dengan baik.
Penelitian ini memberikan kontribusi terhadap pemahaman segmentasi perilaku
pengguna mobile banking dan dapat digunakan sebagai dasar untuk strategi layanan
berbasis waktu yang lebih adaptif dan personal.80 PagesTesis Magiste
Influence of Ease of Transaction and Cashback Promotions on Interest In Using Seabank on D4 Banking And Finance Students Faculty of Vocation University of North Sumatera
No full textDigital banks such as SeaBank have emerged as a form of innovation in the
digitization of banking services offering practical services and attractive
promotions, including cashback programs which are the main attraction among
students. This study aims to determine the effect of ease of transactions and
cashback promotions on interest in using SeaBank for D4 Banking and Finance
students at the Faculty of Vocational, University of North Sumatra. This study uses
quantitative approach by distributing questionnaires to 68 respondents. The data
was analyzed using multiple linear regression. The t test results show that ease of
transaction has a significant effect on interest in use with a significance value of
0.033, and cashback promotions have a more dominant influence with a
significance value 0,000. The F test also shows a significant simultaneous effect
with a significance value of 0.000, with a coefficient determination (R²) of 0.370,
which means that the two variables explain 37% of the variation in interest in using
SeaBank. These results indicate that ease of access and
direct benefits from promotion are important factors in attracting students' interest
in digital banking services.102 PagesKertas Karya Diplom
Detection of Ammonia Gas (NH3) in Paddy Fields Based on The Internet of Things (IoT) and Its Impact on Public Health in The Pangkalan Susu
No full textWhen nitrogen fertilizer is used in agricultural activities, ammonia gas (NH3) emissions increase, which has a negative impact on the health of the surrounding community. The purpose of this research is to develop an NH3 monitoring system in Pangkalan Susu Village, North Sumatra, that uses the Internet of Things (IoT). To send data to the Blynk application in real-time, this system uses an MQ-135 sensor connected to an Arduino Mega 2560 and a SIM800L GSM module. After calibration was performed with a standard AR8500 device, an average accuracy of 97.48% was obtained. The measurement results showed that the NH3 concentration reached 5.25 ppm, exceeding the threshold of 2.00 ppm. In addition, a public health survey was conducted by sending questionnaires to two groups of respondents: farmers and non-farmers. The analysis results showed that exposure to NH3 caused significant health symptoms. The system also serves as an early monitoring tool to detect air pollution and support efforts to reduce and protect public health in rice fields160 PagesSkripsi Sarjan
Risk The occurrence of Nonarteritic Optic Ischemia Neuropathy in Metabolic Syndrome
No full textBackground : Non arteritic anterior ischemic optic neuropathy (NAION) is a condition that causes ischemia and infarction of the optic nerve due to impaired blood supply. Systemic comorbidities such as diabetes, hypertension, and ocular factors such as small optic discs, as well as acute perioperative blood loss are common predisposing factors for NAION. This study aims to determine the risk of metabolic syndrome to metabolic syndrome.
Methods: This is an observational analytical study with a cross sectional data collection method involving 40 people with metabolic syndrome conducted at the Eye Health Polyclinic and Endocrine Polyclinic. Data collection was done through medical records, visual acuity examination, visual field, Farnsworth Munsell D 15, anterior and posterior segments, Optical Coherence Tomography (OCT).
Results: 40 people participated in this study. Distribution of gender show that of 2 cases of NAION in men and in women there are 2 cases. Pressure blood systolic in NAION is more high (146.25 ± 7.5 mmHg). Blood sugar levels fasting in the NAION group has average of 118.75 ± 15.22 mg/dL. Triglyceride levels in the NAION group were 160 ± 29.44 mg/dL. The index mass body mass index (BMI) in the NAION group had average of 25.93 ± 1.69 kg/m²
Conclusion : there is no factors that are significant relate with NAION occurrence in this studies. Age > 65 years and high blood pressure showed trend improvement risk of NAION, but this studies need sample more big for ensure significance clinically. Higher triglyceride levels high in the NAION support suspicion role dyslipidemia in pathophysiology of NAION98 PagesTesis Magiste
Personification of Nations in "Hetalia: World Stars"
No full textThis study discusses the personification of nations in Hetalia: World Stars anime. The personification of nations illustrates national identity by portraying countries as living individuals who possess specific traits and behaviors like humans. This approach is used to make it easier for viewers to understand the concept of national identity through each character. The purpose of this study is to identify the meaning of national personification and to understand the relationships between nations as depicted in Hetalia: World Stars anime, using a literary perspective. The data in this research consist of dialogues, behaviors, and interactions between characters taken from the anime. The method used is descriptive qualitative, with data collection techniques including documentation and observation, which were then analyzed contextually. The results show that the personification of nations in Hetalia: World Stars anime represents the distinctive characteristics of each country, such as Italy’s relaxed nature, Germany’s discipline, Hungary’s empathy, Prussia’s confidence, Austria’s orderliness, France’s sarcasm, America’s arrogance, Slovakia’s consistency, England’s visionary outlook, Russia’s manipulative nature, until Japan’s politeness. The relationships between nations reflect dynamics of cooperation, cultural value conflicts, and political tensions between countries.109 PagesSkripsi Sarjan
In Vitro Test of Ethanol Extract of Attarasa Leaves (Litsea cubeba (Lour.) Pers.) On CT-26 Cell Line
No full textBackground: Colorectal cancer, a prevalent global malignancy, faces treatment challenges due to drug resistance and toxicity. Litsea cubeba, a plant with bioactive compounds, exhibits antibacterial, anti-inflammatory, and antioxidant properties. Emerging evidence suggests its anticancer potential via modulation of the PI3K/AKT pathway, a key driver of cancer progression.
Aim: This study evaluates Litsea cubeba leaves ethanol extract’s phytochemical composition, characterization, and anticancer effects on colorectal cancer, focusing on PI3K/AKT pathway modulation.
Methods: Litsea cubeba leaves from Sidikalang, Indonesia, were authenticated and analyzed macroscopically (leaf morphology, aroma) and microscopically (stomata, vascular bundles). Physicochemical and phytochemical screenings identified bioactive compounds. Ethanolic extract was prepared via maceration and tested for cytotoxicity against CT-26 cells. PI3K/AKT protein expression was assessed via Western Blot.
Results: The extract contained alkaloids, flavonoids, saponins, and tannins. Physicochemical parameters met quality standards. Cytotoxicity assays showed dose-dependent cell death (IC50: 44.52 µg/mL). Western Blot confirmed reduced PI3K/AKT protein levels, indicating pathway suppression. Microscopic analysis revealed anomocytic stomata, oil cells, and vascular structures.
Conclusion: Litsea cubeba leaf extract demonstrates moderate cytotoxicity and PI3K/AKT pathway inhibition in colorectal cancer cells, supporting its potential as a natural therapeutic. Further research is needed to validate efficacy and safety for clinical translation.96 PagesSkripsi Sarjan