Abstract:Abstract : Equitable immunization coverage is a key indicator of successful public health programs, particularly for vulnerable age groups such as toddlers. In Asahan Regency, however, vaccination distribution among districts…
ricts varies significantly and warrants further analysis. This study applies the Self Organizing Map (SOM) method to cluster toddler vaccination data from 25 districts using variables such as total number of toddlers, total vaccination coverage, and gender-specific vaccination percentages. The data was normalized using Z-Score standardization and processed through a SOM grid with 3x1 neurons. The clustering results identified three distinct categories: high, medium, and low coverage, each with unique characteristics. High-coverage clusters were dominated by districts with perfect vaccination rates (100%), while low clusters revealed gender disparities. This study contributes to the visualization and analysis of immunization trends and serves as a reference for more targeted health intervention planning.
Â
Keyword: clustering; self organizing map; SOM; toddlers; vaccination.
Â
Abstrak : Cakupan imunisasi yang merata merupakan salah satu indikator keberhasilan program kesehatan masyarakat, khususnya bagi kelompok usia rentan seperti balita. Namun, di Kabupaten Asahan, distribusi vaksinasi balita menunjukkan ketimpangan antar kecamatan yang perlu dianalisis lebih lanjut. Penelitian ini mengimplementasikan metode Self Organizing Map (SOM) untuk mengelompokkan data vaksinasi balita dari 25 kecamatan berdasarkan variabel jumlah total balita, persentase vaksinasi total, serta vaksinasi berdasarkan jenis kelamin. Data dinormalisasi menggunakan metode Z-Score dan diproses dalam jaringan SOM dengan struktur 3x1 neuron. Hasil klasterisasi menghasilkan tiga kategori cakupan: tinggi, sedang, dan rendah, dengan karakteristik yang berbeda-beda. Klaster tinggi didominasi oleh kecamatan dengan cakupan vaksinasi sempurna (100%), sedangkan klaster rendah menunjukkan adanya kesenjangan cakupan antara balita laki-laki dan perempuan. Penelitian ini memberikan kontribusi dalam visualisasi dan analisis tren imunisasi serta dapat digunakan sebagai acuan dalam perencanaan intervensi kesehatan secara lebih terarah.
Â
Kata kunci: balita; clustering; self organizing map; SOM; vaksinasi
Abstract:Abstract: Determining the optimal routing protocol is crucial in designing a computer network, especially for medium-scale networks that require high efficiency and stability. This study aims to analyze and compare the performance…
erformance of the two most commonly used interior routing protocols, namely OSPF (Open Shortest Path First) and EIGRP (Enhanced Interior Gateway Routing Protocol). Testing was carried out through network simulation using intermediate topology with evaluation parameters including convergence time, bandwidth utilization, throughput level, delay, dan packet loss. The results of the study showed that EIGRP consistently outperforms in delay, jitter, packet loss, throughput, and convergence time, in both star and mesh topologies, while OSPF has stable performance but is slightly slower in the convergence process and processing alternative paths. Thus, the selection of the appropriate routing protocol must be tailored to the specific needs of the network. such as priority over convergence speed or resource efficiency. This study provides a technical overview for network designers in determining the optimal routing protocol for medium-scale computer networks.
Â
Keywords: Routing; OSPF; EIGRP; Throughput; Convergence.
Abstrak: Penentuan protokol routing yang optimal sangat krusial dalam perancangan jaringan komputer, terutama untuk jaringan berskala menengah yang membutuhkan efisiensi dan stabilitas tinggi. Penelitian ini bertujuan untuk menganalisis dan membandingkan kinerja dua protokol routing interior yang paling umum digunakan, yaitu OSPF (Open Shortest Path First) dan EIGRP (Enhanced Interior Gateway Routing Protocol). Pengujian dilakukan melalui simulasi jaringan menggunakan topologi menengah dengan parameter evaluasi meliputi waktu konvergensi, pemanfaatan bandwidth, tingkat throughput, delay, dan packet loss. Hasil penelitian menjukan bahwa EIGRP unggul secara konsisten dalam delay, jitter, packet loss, throughput, dan waktu konvergensi, baik di topologi star maupun mesh, sedangkan OSPF memiliki performa yang stabil namun sedikit lebih lambat dalam proses konvergensi dan pemrosesan jalur alternatif. Dengan demikian, pemilihan protokol routing yang tepat harus disesuaikan dengan kebutuhan spesifik jaringan, seperti prioritas terhadap kecepatan konvergensi atau efisiensi sumber daya. Studi ini memberikan gambaran teknis bagi perancang jaringan dalam menentukan protokol routing yang optimal untuk jaringan komputer skala menengah.
Â
Kata kunci: Routing; OSPF; EIGRP; Troughput; Konvergensi.
Abstract:Abstract: Micro, Small, and Medium Enterprises (MSMEs) in Asahan Regency have very heterogeneous characteristics, both in terms of business sectors, capital size, and income achievement. This study aims to cluster MSMEs…
using the K-Prototypes algorithm to group business actors into uniform clusters based on numerical and categorical characteristics. The methodology used includes the pre-processing stage, variable transformation, determining the optimal number of clusters using the elbow method, and implementing the K-Prototypes algorithm. The results of the study showed that five main clusters were successfully formed, each showing a different pattern in terms of capital, net income, and dominant business sector. Data visualization and exploration (EDA) also strengthened the understanding of the cluster structure that was formed. The cluster with the highest capital and income is dominated by the medium-scale trade sector, while the cluster with the lowest capital and income is identical to micro MSMEs in the culinary and service sectors. These findings prove that the K-Prototypes algorithm is effective in identifying MSME segmentation in a more structured manner and can be the basis for designing more targeted MSME development strategies.
Â
Keyword: UMKM; clustering; K-Prototypes; mixed data; segmentation analysis.
Â
Abstrak: Usaha Mikro, Kecil, dan Menengah (UMKM) di Kabupaten Asahan memiliki karakteristik yang sangat heterogen, baik dari sisi sektor usaha, besaran modal, hingga capaian income. Penelitian ini bertujuan untuk melakukan klasterisasi UMKM menggunakan algoritma K-Prototypes guna mengelompokkan pelaku usaha ke dalam klaster-klaster yang seragam berdasarkan karakteristik numerik dan kategorikal. Metodologi yang digunakan mencakup tahap pre-processing, transformasi variabel, penentuan jumlah klaster optimal menggunakan metode elbow, serta implementasi algoritma K-Prototypes. Hasil penelitian menunjukkan bahwa lima klaster utama berhasil dibentuk, masing-masing menunjukkan pola yang berbeda dalam hal modal, income bersih, dan sektor usaha dominan. Visualisasi dan eksplorasi data (EDA) turut memperkuat pemahaman terhadap struktur klaster yang terbentuk. Klaster dengan modal dan income tertinggi didominasi oleh sektor perdagangan skala menengah, sedangkan klaster dengan modal dan income terendah identik dengan UMKM mikro di sektor kuliner dan jasa. Temuan ini membuktikan bahwa algoritma K-Prototypes efektif digunakan untuk mengidentifikasi segmentasi UMKM secara lebih terstruktur dan dapat menjadi dasar dalam merancang strategi pengembangan UMKM yang lebih tepat sasaran.
Â
Kata kunci: UMKM; klasterisasi; K-Prototypes; data campuran; analisis segmentasi
Abstract:Abstract: This study aims to examine the varieties of language used by Indonesian language teachers in classroom interactions at MAS Bahrul Uluum Al-Kamal, Asahan Regency. It investigates the forms, pragmatic functions,…
and contextual factors influencing the teacher's language choices. The research employed a descriptive qualitative approach with a case study design. Data were collected through participatory observation, audio-video recordings, in-depth interviews, and document analysis. Data analysis followed Miles and Huberman’s interactive model, consisting of data reduction, data display, and conclusion drawing. The findings reveal that teachers employed three primary language varieties: formal, consultative, and casual. Formal language was dominant in delivering lessons and explaining linguistic rules, while consultative and casual varieties emerged during discussions and classroom management. Code-switching and code-mixing phenomena were also identified, involving shifts between Indonesian, regional languages (Melayu and Javanese), and Arabic, especially in religious or culturally significant expressions. These shifts were found to be strategic rather than arbitrary, aimed at enhancing meaning, emotional engagement, and cultural relevance. In terms of pragmatic functions, the teacher’s utterances reflected six functions based on Halliday’s taxonomy: representational, heuristic, interactional, personal, instrumental, and imaginative. Factors influencing language choices included students’ ethnic backgrounds, instructional objectives, cultural-religious context, and the teacher's teaching experience. The use of consultative and casual registers positively impacted student participation and comfort, although it raised concerns regarding the imitation of informal styles in formal academic writing. To address this, a reflective strategy—such as formal language recapitulation at the end of lessons—is recommended. Overall, the study highlights that teacher language variety is not merely a linguistic variation but a pedagogical medium laden with social, cultural, and educational implications, playing a crucial role in shaping students' literacy habits and classroom atmosphere.
Â
Keywords: Language Variety; Code-Switching; Classroom Interaction
Abstrak: Penelitian ini bertujuan untuk mengkaji ragam bahasa yang digunakan guru Bahasa Indonesia dalam interaksi kelas di MAS Bahrul Uluum Al-Kamal, Kabupaten Asahan, dengan menelaah bentuk, fungsi pragmatis, serta faktor-faktor yang memengaruhi pemilihannya. Metode penelitian yang digunakan adalah kualitatif deskriptif dengan pendekatan studi kasus. Data dikumpulkan melalui observasi partisipatif, rekaman audio-video, wawancara mendalam, dan studi dokumentasi. Analisis dilakukan menggunakan model interaktif Miles dan Huberman yang meliputi reduksi data, penyajian data, serta penarikan kesimpulan. Hasil penelitian menunjukkan bahwa guru menggunakan tiga ragam bahasa utama, yaitu ragam resmi, konsultatif, dan santai. Ragam resmi dominan dalam pengantar dan penjelasan materi, sedangkan ragam konsultatif dan santai muncul dalam diskusi serta pengelolaan suasana kelas. Fenomena alih kode dan campur kode juga ditemukan dalam bentuk peralihan antara Bahasa Indonesia, bahasa daerah (Melayu dan Jawa), serta bahasa Arab. Penggunaan variasi ini bersifat strategis untuk memperjelas makna, membangun kedekatan afektif, serta menyesuaikan konteks kultural religius pesantren. Dari segi fungsi pragmatis, tuturan guru mencakup enam fungsi menurut klasifikasi Halliday: representasional, heuristik, interaksional, personal, instrumental, dan imajinatif. Faktor-faktor penentu dalam pemilihan ragam meliputi latar etnis siswa, tujuan instruksional, budaya lokal, serta pengalaman mengajar guru. Penggunaan ragam konsultatif dan santai terbukti meningkatkan partisipasi dan kenyamanan siswa, namun menimbulkan risiko imitasi gaya informal dalam konteks akademik formal. Oleh karena itu, disarankan strategi reflektif seperti rekapitulasi dengan ragam baku di akhir sesi pembelajaran. Penelitian ini menyimpulkan bahwa ragam bahasa guru dalam kelas adalah sarana pedagogik yang tidak hanya mencerminkan variasi linguistik, tetapi juga mengonstruksi dinamika sosial, budaya, dan pembentukan habitus literasi siswa.
Â
Kata kunci: Ragam Bahasa; Alih Kode; Interaksi Kelas
Abstract:Abstract: The rapid development of automotive technology has led to a wide variety of car types with increasingly complex features and characteristics. Alongside this advancement, the need for fast and accurate automatic…
classification systems has become essential, particularly in the field of vehicle image recognition. This study aims to implement the Extreme Learning Machine (ELM) method to classify vehicle types based on digital images. The research focuses on classifying three types of vehicles (Bus, Pickup, and SUV), using a dataset of 10 images per class. The test results show that ELM is capable of classifying vehicle types with competitive accuracy compared to conventional methods, while offering significantly more efficient computation time. This research demonstrates the potential of ELM as a practical solution to support intelligent vehicle recognition systems in the modern automotive industry. The classification achieved an accuracy rate of 91.67%.
Â
Keywords: Extreme Learning Machine (ELM), Vehicle Type Classification, Image Processing, Pattern Recognition.
Â
Abstrak: Perkembangan teknologi otomotif yang pesat telah menghasilkan beragam jenis mobil dengan fitur dan karakteristik yang semakin kompleks. Seiring dengan kemajuan ini, kebutuhan akan sistem klasifikasi otomatis yang cepat dan akurat menjadi sangat penting, terutama dalam bidang pengenalan citra kendaraan. Tujuan penelitian ini untuk mengimplementasikan metode Extreme Learning Machine (ELM) dalam mengklasikasi jenis mobil berdasarkan citra digital. Penelitian ini mengklasifikasikan 3 jenis mobil (BUS, Pickup dan SUV) dengan dataset masing masing 10 gambar. Hasil pengujian menunjukkan bahwa ELM mampu melakukan klasifikasi jenis mobil dengan akurasi yang kompetitif dibandingkan metode konvensional, dengan waktu komputasi yang jauh lebih efisien. Penelitian ini menunjukkan potensi ELM sebagai solusi praktis dalam mendukung sistem cerdas berbasis pengenalan kendaraan dalam dunia otomotif modern. Hasil yang didapat dari penelitian ini adalah sebesar 91,67% untuk tingkat akurasinya.
Â
Kata kunci: Extreme Learning Machine (ELM), Klasifikasi Jenis Kendaraan, Pemrosesan Citra, Pengenalan Pola.
Abstract:Abstract: This study focuses on sentiment analysis of user reviews for the Shopee Partner and Tokopedia Seller applications on Google Play Store using the Naïve Bayes Classifier method. A total of 20,000 user reviews were…
were collected, with 10,000 reviews for each application. The research process included data collection, sentiment labeling, data cleaning, text preprocessing, and classification using Multinomial Naïve Bayes. The analysis results indicate that Shopee Partner reviews were dominated by negative sentiments, achieving an accuracy of 84%, precision of 87%, recall of 92%, and an f1-score of 90%. Conversely, Tokopedia Seller reviews were primarily positive, with an accuracy of 85%, precision of 92%, recall of 83%, and an f1-score of 87%. Data visualization revealed that Shopee Partner had 6,455 negative reviews and 2,324 positive reviews, while Tokopedia Seller recorded 3,728 negative reviews and 5,441 positive reviews. This study concludes that Tokopedia Seller better meets user expectations compared to Shopee Partner. Furthermore, it reaffirms the effectiveness of the Naïve Bayes Classifier as a tool for sentiment analysis.
Â
Keyword: Sentiment Analysis, Shopee Partner, Tokopedia Seller, Naïve Bayes Classifier.
Â
Â
Abstrak: Studi ini berfokus pada analisis sentimen ulasan pengguna terhadap aplikasi Shopee Partner dan Tokopedia Seller di Google Play Store dengan menggunakan metode Naïve Bayes Classifier. Sebanyak 20.000 ulasan pengguna dikumpulkan, masing-masing 10.000 ulasan untuk setiap aplikasi. Proses penelitian mencakup pengumpulan data, pelabelan sentimen, pembersihan data, praproses teks, dan klasifikasi menggunakan Multinomial Naïve Bayes. Hasil analisis menunjukkan bahwa ulasan untuk Shopee Partner didominasi oleh sentimen negatif, dengan akurasi sebesar 84%, presisi 87%, recall 92%, dan f1-score 90%. Sebaliknya, ulasan Tokopedia Seller didominasi oleh sentimen positif, dengan akurasi 85%, presisi 92%, recall 83%, dan f1-score 87%. Visualisasi data menunjukkan bahwa Shopee Partner memiliki 6.455 ulasan negatif dan 2.324 ulasan positif, sedangkan Tokopedia Seller mencatat 3.728 ulasan negatif dan 5.441 ulasan positif. Studi ini menyimpulkan bahwa Tokopedia Seller lebih mampu memenuhi harapan pengguna dibandingkan Shopee Partner. Selain itu, penelitian ini menegaskan efektivitas metode Naïve Bayes Classifier sebagai alat analisis sentimen.
Â
Kata kunci: Analisis Sentimen, Shopee Partner, Tokopedia Seller, Naïve Bayes Classifier.
Abstract:This istudy aimed to explore how marketing communications are implemented at the Fairfield by Marriott Bali Legian Hotel and to analyze how these efforts influence the hotel's brand image. The research used a quantitative…
e method through simple linear regression analysis, supported by a qualitative approach involving structured interviews. The findings show that the hotel has implemented its marketing communications strategy very effectively. In particular, the use of Instagram has generated strong engagement from followers on the official account, @fairfieldbalilegian. This is further supported by an increase in the account’s performance metrics, as measured using the website phlanx.com. Data analysis reveals that marketing communication plays i a significant i role i in enhancing i the hotel’s brand image. The simple regression analysis produced a constant value of 13.420 and a regression coefficient of 0.777, indicating that for every one-unit increase in marketing communication efforts, the brand image improves by 0.777 units. Furthermore, the coefficient of determination (R²) was found to be 0.794. This means that 79.4% of the hotel’s brand image is influenced by marketing communication activities, while the remaining i 20.6% is affected by other i factors
Abstract:The digital economy transformation challenges the relevance of the Permanent Establishment principle in Indonesia’s tax system. As a response, the Significant Economic Presence (SEP) concept was introduced to tax foreign…
foreign digital entities without physical presence. This study aims to examine the conceptual and legal challenges in implementing SEP in Indonesia. The method used is normative-juridical with statutory and conceptual approaches. The findings reveal that SEP implementation faces legal uncertainty, conflicts between domestic and international norms, and lacks clear technical parameters. While SEP holds potential for promoting fiscal justice, its ambiguous legal basis risks creating discrimination unless supported by regulatory reform and international cooperation. This study highlights the urgency of harmonizing tax laws and strengthening institutional capacity to meet global digital taxation challenges.
Abstract:Abstract: Air pollution has become a serious problem in densely populated urban areas such as DKI Jakarta. To mitigate its negative impacts, an accurate air pollution prediction system is necessary. This study compares the…
he performance of two deep learning models, Long Short-Term Memory (LSTM) and Bidirectional Long Short-Term Memory (BiLSTM), in predicting PM10 concentration using air quality data from DKI Jakarta between 2016 and 2019. The research process includes data collection and preprocessing, model training, and model evaluation. Both models were tested with various parameters such as the number of hidden neurons, dropout rate, epochs, and batch size. The results consistently show that BiLSTM outperforms LSTM, achieving lower Root Mean Square Error (RMSE) values across 54 testing scenarios. The best BiLSTM configuration, with 64 hidden neurons, 0.2 dropout rate, 50 epochs, and batch size 16, yielded an RMSE of 9.311401. Meanwhile, the best LSTM configuration, with 128 hidden neurons, 0.1 dropout rate, 100 epochs, and batch size 16, produced an RMSE of 9.330554. The advantage of BiLSTM lies in its ability to process data bidirectionally, making it more effective in capturing temporal patterns for air pollution prediction compared to LSTM.
Â
Keywords: air pollution prediction, pollutant, deep learning, LSTM, BiLSTM
Â
Abstrak: Pencemaran udara menjadi masalah serius di wilayah perkotaan padat seperti DKI Jakarta. Untuk mengurangi dampak negatifnya, diperlukan sistem prediksi polusi udara yang akurat. Penelitian ini membandingkan performa dua model deep learning, Long Short-Term Memory (LSTM) dan Bidirectional Long Short-Term Memory (BiLSTM), dalam memprediksi konsentrasi PM10 menggunakan data kualitas udara DKI Jakarta tahun 2016-2019. Proses penelitian mencakup pengumpulan dan praproses data, pelatihan model, serta evaluasi model. Kedua model diuji dengan berbagai parameter seperti jumlah hidden neuron, dropout rate, epoch, dan batch size. Hasil menunjukkan BiLSTM lebih unggul secara konsisten dengan nilai Root Mean Square Error (RMSE) lebih rendah melalui 54 skenario pengujian. Konfigurasi terbaik BiLSTM menggunakan 64 hidden neuron, dropout rate 0.2, 50 epoch, dan batch size 16 menghasilkan RMSE 9.311401. Sedangkan konfigurasi LSTM terbaik pada 128 hidden neuron, dropout rate 0.1, 100 epoch, dan batch size 16 menghasilkan RMSE 9.330554. Keunggulan BiLSTM terletak pada kemampuannya memproses data dua arah, sehingga lebih efektif dalam menangkap pola temporal untuk prediksi polusi udara dibandingkan LSTM.Â
Â
Kata kunci: prediksi polusi udara, polutan, deep learning, LSTM, BiLSTM
Abstract:Abstract: This study aims to analyze the distribution of teaching load of teachers at SMKN 1 Stabat using the K-Means Clustering method. The main problem faced is the imbalance in the number of teaching hours between teachers…
chers which can affect the effectiveness of teaching and learning activities. The data used is the total weekly teaching hours of each teacher, which is taken from the document on the distribution of teaching tasks for the even semester of the 2024/2025 academic year. The K-Means method is used to group teachers into three categories of teaching load: light, medium, and heavy. The grouping process is carried out by determining the number of clusters (K = 3), initializing the centroid, calculating the distance of each data to the centroid, and updating the position of the centroid until the results are stable. The final results show that most teachers are included in the medium load cluster, while a small number are in the light and heavy categories. This shows that the distribution of the teaching load is not yet completely even. The application of K-Means has been proven to be able to provide an analytical picture of the distribution of teacher work, as well as support data-based decision making in education management.
Â
Keywords: K-Means Clustering, Teaching Load, Teacher Schedule, Data Clustering, SMKN 1 Stabat
Abstrak: Penelitian ini bertujuan untuk menganalisis pembagian beban mengajar guru di SMKN 1 Stabat dengan menggunakan metode K-Means Clustering. Masalah utama yang dihadapi adalah ketidakseimbangan jumlah jam mengajar antar guru yang dapat memengaruhi efektivitas kegiatan belajar-mengajar. Data yang digunakan berupa total jam mengajar mingguan dari setiap guru, yang diambil dari dokumen pembagian tugas mengajar semester genap tahun ajaran 2024/2025. Metode K-Means digunakan untuk mengelompokkan guru ke dalam tiga kategori beban mengajar: ringan, sedang, dan berat. Proses pengelompokan dilakukan dengan menentukan jumlah klaster (K=3), menginisialisasi centroid, menghitung jarak masing-masing data ke centroid, serta memperbarui posisi centroid hingga hasil stabil. Hasil akhir menunjukkan bahwa sebagian besar guru tergolong dalam klaster beban sedang, sementara sebagian kecil masuk kategori ringan dan berat. Hal ini menunjukkan bahwa distribusi beban mengajar belum sepenuhnya merata. Penerapan K-Means terbukti mampu memberikan gambaran analitis terhadap distribusi kerja guru, serta mendukung pengambilan keputusan berbasis data dalam manajemen pendidikan.
Â
Kata kunci: K-Means Clustering, beban mengajar, jadwal guru, pengelompokan data, SMKN 1 Stabat