Claim Missing Document
Check
Articles

Found 39 Documents
Search

Implementation Of Agglomerative Clustering Method On Mapping Crime-Prone Areas Of Webgis-Based Lhokseumawe City Case Study Of Lhokseumawe Prosecutor's Office Teuku Ibrar Faturrahman Ibrar; Safwandi Safwandi; Zahratul Fitri Zahratul Fitri
Proceedings of International Conference on Multidisciplinary Engineering (ICOMDEN) Vol. 2 (2024): Proceedings of International Conference on Multidisciplinary Engineering (ICOMDEN)
Publisher : Faculty of Engineering, Malikussaleh University

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

The application of the Agglomerative Hierarchical Clustering method was carried out for mapping the Lhokseumawe City area, with a focus on sub-districts grouped by village and their crime rates. The types of crimes analyzed include drugs, oharda (violations of public order and security), and kamtibum (public order and security). The data used came from the Prosecutor's Office and was taken through the Department of Law, covering various crimes that are very detrimental to society. By utilizing Geographic Information System (GIS) technology, this system can provide clear visual information about the location of criminal incidents and the types of crimes that occur in each village. This clustering process allows for the grouping of villages that tend to have high crime rates, thus helping to identify areas that require more attention in law enforcement. The application of this clustering is not new, because previously many researchers and scientists have applied similar methods, but with different case studies. In this context, clustering helps provide more detailed insights into the distribution of crime at the village level, allowing for more focused and targeted prevention efforts.
Implementation Of The Support Vector Machine Method In Determining The Best Quality Of Sap Azhari Putra Sayani; Safwandi; Fajriana
Proceedings of International Conference on Multidisciplinary Engineering (ICOMDEN) Vol. 2 (2024): Proceedings of International Conference on Multidisciplinary Engineering (ICOMDEN)
Publisher : Faculty of Engineering, Malikussaleh University

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

Rubber trees (Hevea brasiliensis) are the main source of natural rubber and play an important role in Indonesia's industry. Determining the quality of rubber sap is a challenge for companies, as traditional manual processes are time-consuming and prone to human error. PT Poly Kencana Raya, a company in Besitang, North Sumatra, currently still uses conventional methods in determining the quality of rubber latex it produces. This research aims to design a web-based system with the application of the Support Vector Machine (SVM) method to facilitate the determination of rubber latex quality. SVM was chosen as a classification method because of its ability to determine the optimal hyperplane that can separate data from two different classes, namely feasible and unfit. The built system utilizes the main criteria such as tree age, tapping time, moisture content, color, and texture in determining the quality of the sap. Implementation. This study used 120 samples of test data, with accurate prediction results on 111 data, resulting in an accuracy rate of 92.5%. This decision support system is expected to increase efficiency and accuracy in rubber sap selection and support the development of rubber production quality in Indonesia. This research also opens up opportunities for further development by adding other classification methods for accuracy comparison or adding training data to optimize prediction results. Keywords: Rubber Trees, Support Vector Machine, Data Mining
Implementasi Data Mining Penentuan Daya Pelanggan Baru untuk Klasifikasi Subsidi dan Non Subsidi di Wilayah PLN Kota Lhokseumawe Andik Bintoro; Safwandi Safwandi
Sisfo: Jurnal Ilmiah Sistem Informasi Vol. 2 No. 2 (2018): Sisfo: Jurnal Ilmiah Sistem Informasi, Oktober 2018
Publisher : Universitas Malikussaleh

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29103/sisfo.v2i2.1011

Abstract

Penerapan model data mining dalam klasifiaksi K-Nearest Neighbor (KNN) dalampenentuan Daya Pelanggan Baru untuk Klasifikasi Subsidi dan Non Subsidi diWilayah PLN Kota Lhokseumawe sangatlah perlu untuk dilakukan. Hal ini untukmengetahui pelanggan PLN dalam klasifikasi pengelompokkan berdasarkandengan variable yang digunakan. Pelanggan baru PLN yang ingin mengetahuiberapa klasifikasi pengelompokkan listrik untuk pasang baru dapat menggunakan aplikasi ini dalam penentuan besaran daya yang digunakan untuk pasang baru. Variable yang digunakan adalah untuk K1 adalah pendapatan keluarga secara keseluruhan, K2 adalah pekerjaan orang tua, K3 adalah pekerjaan orang tua, K4 adalah luas rumah. Dari variable ini akan dimasukkan kedalam sistem dengan sub kriteria yang akan dimasukkan. Adapun Golongan tarif daya listrik subsidi terbagi menjadi dua yaitu Tarif R-1/450 VA untuk penggolongan K1 dan golongan R-1/900 VA Subsidi Tarif R-1/586 kWh untuk penggolongan K2, sedangkan Golongan K3 non subsidi Tarif R-1/1352 VA untuk rumah tangga mampu. Tujuan dari penelitian ini adalah memudahkan pelanggan baru dalam penentuan golongan yang diberikan oleh pihak PLN kota lhokseumawe dan pelanggan mengetahui besaran yang akan digunakan. Sehingga pelanggan kota lhokseumawe dapat melihat besaran jumlah daya yang diberikan. Model klasifikasi KNN menggunakan data training untuk di uji coba dengan data pengujian untuk pelanggan baru dan hasilnya dari nilai terdekat dari data training yang telah diujikan. Dalam penelitian ini data training yang digunakan adalah 30 data training, hal ini berguna untuk nilai pengujian lebih akurat dan hasil yang diharapkan menjadi lebih baik. Untuk pengujian V1 adalah dengan nilai 3, untuk kriteria V2 adalah 3, dan nilai variabel V=3 adalah 2 dan nilai variabel terakhir V4 adalah 1. Selanjutnya untuk hasil pengujian dari pelanggan baru adalah termasuk kedalam golongan subsidi R-1/900 VA Subsidi Tarif R-1/586 kWh yang telah dilihat dari hasil training. Hasil dari aplikasi ini dapat menampilkan besaran daya yang digunakan untuk pelanggan baru berdasarkan nilai pengujian yang dilakukan.Kata kunci : Klasifikasi, Daya Listrik ,KNN
IoT-Based LPG Gas Leak Monitoring System with Automatic Alarms Muhammad Hafizal; Safwandi Safwandi; Kurniawati Kurniawati; Muchlis Abd Muthalib; Said Fadlan Anshari
SISTEMASI Vol 15, No 6 (2026): Sistemasi: Jurnal Sistem Informasi
Publisher : Universitas Islam Indragiri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32520/stmsi.v15i6.6348

Abstract

Liquefied Petroleum Gas (LPG) leaks are one of the leading causes of fires in households and LPG distribution facilities. Limited public awareness and delays in detecting gas leaks can significantly increase the risk of property damage and endanger human safety. Therefore, this study aims to design and develop an Internet of Things (IoT)-based monitoring system for 3 kg LPG gas leaks that provides automatic, real-time early warning notifications. The proposed system employs an MQ-2 gas sensor to detect LPG concentration, a NodeMCU ESP8266 as the main controller and data communication module, and a buzzer, LED, and LCD as local warning indicators. A threshold-based method is implemented to classify safe and hazardous conditions according to predefined gas concentration limits. In addition, the system is integrated with remote notifications via Telegram, enabling users to receive alerts even when they are away from the monitored location. Experimental results demonstrate that the proposed system can accurately detect LPG gas leaks under various operating conditions and automatically send warning notifications whenever the gas concentration exceeds the predefined threshold. These findings indicate that the system is effective, responsive, and has strong potential to enhance user safety in the use of LPG.
SISTEM REKOMENDASI DAN PEMETAAN DESTINASI WISATA ALAM TERBAIK DI BERASTAGI MENGGUNAKAN METODE TECHNIQUE FOR ORDER PREFERENCE BY SIMILARITY TO IDEAL SOLUTION (TOPSIS) BERBASIS WEB Nadya Syahfitri Damanik; Safwandi; Fajriana
Rabit : Jurnal Teknologi dan Sistem Informasi Univrab Vol 10 No 2 (2025): Juli
Publisher : LPPM Universitas Abdurrab

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36341/rabit.v10i2.6044

Abstract

Berastagi is known as one of the favorite natural tourist destinations in North Sumatra that offers a variety of interesting tourist attractions. However, the many choices often make it difficult for tourists to determine the location that suits their preferences. This study aims to develop a web-based recommendation and mapping system for the best natural tourist destinations in Berastagi by implementing the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method. This system considers six main criteria, namely ease of vehicle access, consumption costs, accommodation costs, security levels, and the availability of toilets and parking areas. Tourist locations are visualized through interactive maps based on Geographic Information Systems (GIS). Data were obtained from field surveys, interviews with visitors and managers, and location mapping using Google Maps. This system is specifically designed for three main user groups: independent tourists who want to plan their trips independently, local travel agencies who need data-based guidance to compile tour packages, and the district tourism office as a basis for decision making in promoting and developing regional tourism. System validation was carried out by comparing the recommendations given by the system with the actual choices of 20 respondents. From the results of this test, the system showed a match rate of 85%, with a p-value of 0.021. This figure shows that the recommendations from the system are quite in accordance with user preferences and have statistically significant results. Based on the analysis using the TOPSIS method, the five natural tourist destinations with the highest values ​​recommended by the system are: Lau Sidebuk-Debuk Hot Springs (0.694387), Mount Sibayak (0.648177), Sipiso-Piso Waterfall (0.644718), Gundaling Hill (0.636776), and Lau Kawar Lake (0.625521). This system is expected to be a guide that helps tourists plan trips more precisely, as well as support digital promotion of leading natural tourist attractions in the Berastagi area.
PENERAPAN PERCEIVED STRESS SCALE DALAM PSYCHOLOGICAL SELF-ASSESSMENT UNTUK MENGUKUR TINGKAT STRES MENGGUNAKAN METODE K-NEAREST NEIGHBORS Nadia Rizatul Riski; Safwandi Safwandi; Said Fadlan Anshari
Rabit : Jurnal Teknologi dan Sistem Informasi Univrab Vol 10 No 2 (2025): Juli
Publisher : LPPM Universitas Abdurrab

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36341/rabit.v10i2.6479

Abstract

Stress is a common psychological issue experienced by university students, particularly in high-pressure academic environments such as engineering faculties. This study aims to develop a digital self-assessment system to measure student stress levels using the Perceived Stress Scale (PSS-10) and the K-Nearest Neighbors (KNN) classification method. Data were collected from 100 respondents from the Faculty of Engineering at Universitas Malikussaleh. The system classifies stress levels into three categories: mild, moderate, and severe. Testing was conducted using 5-fold cross-validation. The evaluation results show an average accuracy of 83%, weighted precision of 79.18%, weighted recall of 84.70%, and weighted F1-score of 80.10%. These findings indicate that the system is capable of providing fairly accurate stress classification and can serve as a useful tool for independent stress detection.
IMPLEMENTASI MAUT PADA SISTEM PENDUKUNG KEPUTUSAN DALAM DISTRIBUSI PERSEDIAAN KANTONG DARAH DI UDD PMI KOTA LHOKSEUMAWE: IMPLEMENTATION OF MAUT IN DECISION SUPPORT SYSTEMS FOR BLOOD BAG DISTRIBUTION AT THE PMI BLOOD TRANSFUSION UNIT IN LHOKSEUMAWE CITY Roza Tarina; Safwandi Safwandi; Yessy Afrillia
Rabit : Jurnal Teknologi dan Sistem Informasi Univrab Vol 10 No 2 (2025): Juli
Publisher : LPPM Universitas Abdurrab

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36341/rabit.v10i2.6520

Abstract

The Blood Donor Unit (UDD) of PMI Lhokseumawe City has a vital role in maintaining the availability and distribution of blood bags to various hospitals. However, the distribution process is faced with the challenge of a mismatch between the number of requests and fluctuating supplies, which risks disrupting medical services and causing waste due to expired blood. This study aims to design and implement a web-based decision support system (SPK) that assists PMI UDD in prioritizing the distribution of blood bags objectively and efficiently. The method used is Multi-Attribute Utility Theory (MAUT), with evaluation of six main criteria: number of uses, number of supplies, number of requests, road density, travel time, and distance. The system was built using PHP and MySQL, and tested using the black-box testing method. The test results show that all system features work validly according to the expected functionality, thus supporting the reliability of the application in the decision-making process. The MAUT calculation results show that RSU Metro Medical Center obtained the highest utility value of 0.5809 (58.09%), placing it as the top priority for distribution. In comparison, RSU Cut Meutia obtained 0.4234 (42.34%), while RSIA Bunda obtained 0.3997 (39.97%), which shows a significant difference in distribution priorities. The implementation of this system is expected to improve distribution effectiveness, optimize inventory allocation, and support more accurate and measurable decision-making at UDD PMI Lhokseumawe City.
Information Technology Governance Audit Using COBIT 5 of DSS Domain (Deliver, Service, And Support) Framework at Malikussaleh University Lhokseumawe Safwandi Safwandi; Muthmainnah Muthmainnah; Misbahul Jannah
Journal of Renewable Energy, Electrical, and Computer Engineering Vol. 2 No. 1 (2022): March 2022
Publisher : Institute for Research and Community Service (LPPM), Universitas Malikussaleh, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29103/jreece.v2i1.6633

Abstract

Information technology is very important for companies or institutions to support the achievement of the company's strategic plans to achieve their goals, vision, and mission. Nowadays, an institution can improve the performance of information technology that goes hand in hand with the development of information technology to produce better technology by auditing information technology governance in the company. The purpose of this study is to analyze IT governance using the COBIT 5 framework in the DSS domain at Malikussaleh University Lhokseumawe. COBIT 5 provides a comprehensive framework that helps companies achieve their goals in corporate governance and IT governance. The framework helps companies create optimum value from IT by maintaining a balance between realizing benefits and optimizing risk and resource usage levels. By conducting an audit of information technology governance in the company, the company can find out whether the information technology that has been operating is in accordance with the business processes and company objectives and convey accurately based on the IT strategy. The results of the information technology governance audit based on COBIT 5 in the DSS Domain, on average are at 2.2 (Manage process) to 2.6 (Established Process).
PENGELOMPOKAN UMKM SEKTOR INDUSTRI MIKRO MENGGUNAKAN METODE K-MEANS CLUSTERING BERBASIS ANDROID DI ACEH UTARA Kamilna kamilna; Safwandi Safwandi; Zahratul Fitri
Djtechno: Jurnal Teknologi Informasi Vol 7, No 2 (2026): Agustus
Publisher : Universitas Dharmawangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.46576/djtechno.v7i2.9670

Abstract

Usaha Mikro, Kecil, dan Menengah (UMKM) merupakan salah satu sektor yang berperan penting dalam mendukung perekonomian daerah. Di Kabupaten Aceh Utara, jumlah UMKM mikro sektor industri yang cukup besar menyebabkan proses pengelompokan data secara manual menjadi kurang efektif dalam mendukung penyusunan program pembinaan. Penelitian ini bertujuan mengimplementasikan metode K-Means Clustering ke dalam aplikasi berbasis Android untuk mengelompokkan UMKM mikro sektor industri berdasarkan karakteristik aset dan omset. Data penelitian diperoleh dari Dinas Perdagangan, Perindustrian, Koperasi, dan Usaha Kecil Menengah Kabupaten Aceh Utara sebanyak 1.151 data yang telah melalui proses preprocessing. Pengujian dilakukan dalam dua tahap, yaitu pengujian algoritma menggunakan sampel sebanyak 100 data UMKM serta implementasi pada aplikasi menggunakan seluruh data penelitian. Hasil pengujian terhadap 100 data menghasilkan Cluster 1 sebanyak 71 UMKM, Cluster 2 sebanyak 19 UMKM, dan Cluster 3 sebanyak 10 UMKM. Implementasi pada aplikasi mencapai kondisi konvergen pada iterasi ke-6 dengan hasil akhir Cluster 1 sebanyak 1.106 UMKM (96,09%), Cluster 2 sebanyak 44 UMKM (3,82%), dan Cluster 3 sebanyak 1 UMKM (0,09%). Hasil penelitian menunjukkan bahwa metode K-Means Clustering mampu mengelompokkan UMKM berdasarkan tingkat kemiripan karakteristik aset dan omset serta berhasil diimplementasikan ke dalam aplikasi berbasis Android. Aplikasi yang dikembangkan diharapkan dapat membantu Dinas Perdagangan, Perindustrian, Koperasi, dan UKM Kabupaten Aceh Utara dalam memperoleh informasi pengelompokan UMKM sebagai dasar penyusunan program pembinaan dan pengambilan keputusan berbasis data
PERBANDINGAN PREDIKSI HARGA SAHAM DENGAN MENGGUNAKAN METODE RECURRENT NEURAL NETWORK DAN LONG SHORT TERM MEMORY Fatia Naura; Safwandi; Rizki Suwanda
Rabit : Jurnal Teknologi dan Sistem Informasi Univrab Vol 11 No 2 (2026): Juli
Publisher : LPPM Universitas Abdurrab

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36341/rabit.v11i2.8345

Abstract

Unpredictable stock price fluctuations encourage the use of artificial intelligence methods based on deep learning. This study compares the performance of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) in predicting the share price of PT Pacific Strategic Financial Tbk. using 346 daily historical data (January 2024–June 2025) sourced from Investing.com. The data collected includes daily stock prices, such as opens, closes, highs, lows, and trading volumes, which will be used to train and test the prediction model. The research stages include pre-processing of data (Min-Max normalization and windowing), model design and training, evaluation using Mean Squared Error (MSE) and Mean Absolute Error (MAE), 5-fold cross-validation, and window size sensitivity analysis. The results showed that RNN was slightly superior to LSTM in prediction accuracy (MSE 0.001417 vs 0.001514; MAE 0.030937 vs 0.031938), inter-fold consistency, and computational efficiency of RNN parameters is only a quarter and memory usage is 1.5 times more efficient than LSTM. In contrast, LSTMs produce predictive patterns that are more visually refined and potentially more suitable for long-term trend analysis. The limited number of data (346 observations) on one issuer is allegedly a factor that limits the theoretical advantages of LSTM gating architecture. These findings imply that RNNs are a more efficient option for short-term stock price predictions with limited data, while LSTMs are more relevant for long-term trend analysis needs on larger volumes of data.