Jurnal Informatika: Jurnal Pengembangan IT
Not a member yet
437 research outputs found
Sort by
Peramalan Permintaan Produk Menggunakan ARIMA Berbasis Data Mining
No full textDemand forecasting is a crucial component of business strategy to anticipate customer demand fluctuations and optimize inventory management. Data mining serves as an important analytical approach to uncover hidden patterns in historical data, enabling the generation of accurate predictions. This study aims to forecast the demand for association-related products at Toko As-Sakinah ’Aisyiyah using the Auto Regressive Integrated Moving Average (ARIMA) method, with Moving Average employed as a baseline comparison model. The dataset consists of monthly sales data aggregated from daily records spanning the period of January 2020 to December 2024, resulting in a total of 60 observations. The research stages followed the CRISP-DM framework, encompassing business understanding, data preparation, modeling, evaluation, and deployment. The analysis results indicate that the ARIMA(1,1,1) model is the most suitable, as it meets residual assumptions and yields lower error values compared to Moving Average. The comparison further confirms that ARIMA is more adaptive to trend patterns and short-term fluctuations. The 2025 demand projection reveals a consistent upward trend from January to December. Based on these findings, it is recommended that the store management gradually increase inventory levels to prevent supply shortages in the futur
Prediksi Stok Barang di Toko Eko Helm Menggunakan Metode Time series Analysis
No full textEko Helm Store located in South Lampung, faces challenges in managing helmet inventory, particularly in determining the optimal stock levels for two categories: affordable and premium helmets. This study aims to forecast helmet stock requirements for the year 2024 using the ARIMA method. Weekly sales data from January to December 2024 were analyzed through stationarity testing using the Augmented Dickey-Fuller (ADF) test and differencing, followed by parameter identification based on ACF and PACF plots. The best-fitting models were identified as ARIMA(2,1,0) for premium helmets, with a Mean Squared Error (MSE) of 24.5101 and an Akaike Information Criterion (AIC) of 249.4062, and ARIMA(1,1,0) for affordable helmets, with an MSE of 32.6102 and an AIC of 250.5381. ARIMA was selected due to its ability to capture trends and seasonal fluctuations more effectively than methods such as moving average or exponential smoothing. The forecasting results estimate a stock requirement of 112 units for affordable helmets and 64 units for premium helmets over the next four weeks. The ARIMA model is integrated into an automated forecasting system that runs scheduled scripts without manual intervention. This system supports timely and precise inventory procurement decisions
Perbandingan Metode KNN dan Naïve Bayes dalam Deteksi Tingkat Stres Berdasarkan Ekspresi Wajah
No full textStress is a feeling in which a person feels under pressure, overwhelmed, and has difficulty in dealing with a problem. Stress can be caused by various factors, such as academic pressure, work, personal problems, or social environment. If not addressed immediately, stress can have adverse effects on an individual's health, such as causing high blood pressure, heart disease, sleep disturbances, and a decreased immune system, which makes a person more vulnerable to various diseases. Therefore, monitoring stress levels is very important to prevent more serious negative impacts. Generally, stress detection is done through consultation with a psychologist, but this method has a subjective nature and requires a lot of time and money. Therefore, this research develops a computer vision-based stress detection system using OpenCV and Dlib, with K-Nearest Neighbors and Naïve Bayes algorithms. The data of 500 samples is divided into 80% training data and 20% test data. Features were extracted, and stress was classified into three levels: low, medium and high. Evaluation using k-fold cross-validation (n_split=10, random_state=42) based on accuracy, precision, recall, and F1-score. The results showed that K-Nearest Neighbors with k=5 excelled with 74% accuracy, 73% precision, 73% recall, and 73% F1-score. Meanwhile, Naïve Bayes only achieved 52% accuracy, 51% precision, 48% recall, and 41% F1-score. This shows that KNN is more effective in stress level classification. However, the accuracy of the model is still limited due to the small amount of training data. Parameter optimization and dataset addition are required to improve the overall system performance
Perbandingan Inisialisasi Bobot Random dan Nguyen-Widrow Pada Backpropagation Dalam Klasifikasi Penyakit Diabetes
No full textDiabetes is a metabolic disorder that occurs when the pancreas is unable to produce adequate amounts of insulin or the body has difficulty in utilizing it optimally. This condition has the potential to cause various health complications. Therefore, early diagnosis of diabetes is very important to reduce the mortality rate due to these complications. Backpropagation Neural Network (BPNN) is an approach in Artificial Neural Network (ANN) that is commonly applied for disease classification, including diabetes. However, the BPNN method has drawbacks, namely its slow convergence rate and the possibility of getting stuck at a local minimum due to random weight initialization. To overcome these problems, this study applies the Nguyen-Widrow weight initialization method to improve the performance of BPNN in diabetes classification. The data source in this study comes from Kaggle, consisting of 768 data with 8 parameters. Model testing was conducted using k-fold cross-validation with K=10, and exploring various numbers of neurons in the hidden layer and learning rate (lr). The results showed that weight initialization using the Nguyen-Widrow method improved the accuracy of BPNN compared to random weight initialization. The best model was obtained with lr 0.001 and 15 neurons in the hidden layer, resulting in an accuracy of 91.23%, higher than the random weight initialization which only reached 89.91%. Thus, the Nguyen-Widrow method is proven effective in improving the performance of BPNN for diabetes classification.Diabetes merupakan gangguan metabolik yang terjadi ketika pankreas tidak mampu menghasilkan insulin dalam jumlah yang memadai atau tubuh mengalami kesulitan dalam memanfaatkannya secara optimal. Kondisi ini berpotensi menimbulkan beragam komplikasi kesehatan. Oleh karena itu, diagnosis dini penyakit diabetes sangat penting untuk menekan angka kematian akibat komplikasi tersebut. Backpropagation Neural Network (BPNN) adalah pendekatan dalam Jaringan Syaraf Tiruan (JST) yang umum diterapkan untuk klasifikasi penyakit, termasuk diabetes. Namun, metode BPNN memiliki kekurangan, yaitu laju konvergensinya yang lambat dan kemungkinan terjebak pada minimum lokal akibat inisialisasi bobot yang dilakukan secara random. Untuk mengatasi permasalahan tersebut, penelitian ini menerapkan metode inisialisasi bobot Nguyen-Widrow guna meningkatkan performa BPNN dalam klasifikasi diabetes. Sumber data dalam penelitian ini berasal dari Kaggle, terdiri dari 768 data dengan 8 parameter. Pengujian model dilakukan menggunakan k-fold cross-validation dengan K=10, serta mengeksplorasi berbagai jumlah neuron dalam hidden layer dan learning rate (lr). Hasil penelitian menunjukkan bahwa inisialisasi bobot menggunakan metode Nguyen-Widrow meningkatkan akurasi BPNN dibandingkan dengan inisialisasi bobot random. Model terbaik diperoleh dengan lr 0,001 dan 15 neuron pada hidden layer, menghasilkan akurasi sebesar 91,23%, lebih tinggi dibandingkan inisialisasi bobot random yang hanya mencapai 89,91%. Dengan demikian, metode Nguyen-Widrow terbukti efektif dalam meningkatkan performa BPNN untuk klasifikasi diabetes
Penerapan Data Mining Dalam Pemberian Kelayakan Kredit Nasabah Pada Badan Usaha Milik Desa Gedong Gincu Dengan Metode Naïve Bayes
No full textBUMDes Gedong Gincu is a rural business that provides savings and loan services to village communities. BUMDes Gedong Gincu in determining customers is still very simple, namely by looking at where the customer comes from and how close they are to previous customers who have applied for credit, so it is not a guarantee whether or not the new customer is eligible to get a loan at BUMDes Gedong Gincu. Another problem that occurs is the lack of thoroughness by BUMDes administrators in assessing customers. The aim of this research is to apply the Naïve Bayes method in determining the credit worthiness of customer applications. The method used in this research is Naïve Bayes with testing using a confusion matrix. Research stages include data selection, data cleaning, data transformation, application of the Naïve Bayes method, testing. The result of this research is an application that can facilitate BUMDes Gedong Gincu in evaluating the feasibility of providing credit. The test results using 186 data using the confusion matrix method, an accuracy of 67% was obtained. However, based on SUS testing by users, they got a result of 82.25. This indicates that this application is good and suitable for use by BUMDes Gedong Gincu
Deteksi Tepi Menggunakan Metode Operator Prewitt dan Kirsch pada Citra Uang Kertas
No full textThe importance of edge detection in image processing, especially on banknote objects, prompted this research to carry out an analysis of two edge detection methods, namely the Prewitt and Kirsch operators. five images of banknotes with different denominations (2000, 5000, 10000, 20000 and 50000) were taken as research objects. The edge detection method is implemented using MATLAB, utilizing both Prewitt and Kirsch operators. Image quality assessment uses PSNR, Histogram and Pixel value parameters. The comparison results show that the Prewitt and Kirsch operators provide optimal edge detection results, producing clear and sharp edges in the banknote image. The edge detection quality assessment was carried out through the PSNR metric, and both showed PSNR values above 30 dB, indicating good quality in terms of clarity and accuracy. Comparison of the Histogram and Pixel values shows that the Kirsch method has a higher Histogram and the Prewitt method has a higher Pixel value
Perancangan Sistem Informasi Kesehatan Siswa Berbasis Web "Student Health Electronic Information System (SHEISys)"
No full textDesign of the Student Health Electronic Information System (SHEISys), a web-based health information system for students, is covered in this paper. The ineffectiveness of overseeing and controlling pupils' health situations at school prompted the development of the system. This study's primary goal is to develop and put into place a system that can combine and digitize student health data. The System Development Life Cycle (SDLC) with the Waterfall model is the methodology used in the system development process. It covers a number of steps, including needs analysis, system design using UML diagrams like the Use Case Diagram, Activity Diagram, and Class Diagram, implementation using PHP, Laravel, and MySQL, and black box testing. The results demonstrate that SHEISys effectively satisfies all of the functional criteria, including major data management, health service recording, user authentication, report generation, and student health statistics. All features operate as anticipated, producing reliable and consistent results, according to the findings of the black box testing. Additionally, the system has dashboard functions, exam sites, and reports that facilitate information openness and parental participation in their children's health monitoring. According to the study's findings, SHEISys has the potential to develop into a viable and successful information technology solution that supports the School Health Program (UKS) and enhances student health services at school
Text Summarization Umpan Balik Pengguna Website SiBayar Pondok Pesantren Sabilurrosyad dengan Metode Bi-LSTM
No full textThe SiBayar website is being developed by the Sabilurrosyad Islamic Boarding School to facilitate its administration and management. To improve the functionality of the website, user feedback is needed on the existing features. However, managing and analyzing a large amount of user feedback manually can be a very time-consuming process. Therefore, an automated approach such as text summarization is needed to summarize and analyze the data. This study aims to generate an automated summary of user feedback on the SiBayar website of the Sabilurrosyad Islamic Boarding School using the Bi-Directional Long Short-Term Memory (Bi-LSTM) method, focusing on identifying the best parameters through hyperparameter tuning and evaluating the accuracy in full. The results of the hyperparameter tuning test show that the configuration that provides the best performance is the one using the Nadam algorithm optimization, the number of layers 1 and batch size 1, and the variational dropout with a dropout rate of 0.5. The model summary quality evaluation was performed using the ROUGE metric which showed that the Bi-LSTM model achieved a ROUGE-1 score of 0.6221, a ROUGE-2 score of 0.5462, and a ROUGE-L score of 0.660. Overall, Bi-LSTM model in this study has good performance in summarizing text, but the suitability of word pairs and sequences still needs to be improved for more optimal results
Implementasi Algoritma Binary Space Partitioning Untuk Procedural Map Generation Dalam Gim
No full textThe popularity of games as an interactive entertainment medium continues to grow, with 2D maps playing a vital role in enhancing user experience. Manual map creation is time-intensive, particularly as game worlds become increasingly complex. Procedural Content Generation (PCG) offers a solution by automating map creation, improving replayability, and reducing designer workload. This research explores the use of the Binary Space Partitioning (BSP) algorithm for procedural dungeon map generation, incorporating random connections between rooms to create more exploratory and dynamic maps. The process includes three stages: developing a dungeon map generator, implementing BSP with random room connectors, and validating the generated maps to ensure navigability. Space Syntax analysis, including Visibility Graph Analysis (VGA) and Axial Line Analysis, is applied to evaluate the quality of the maps in terms of connectivity, visibility, and integration. Results show that BSP-generated maps with random connections offer dynamic layouts, while Space Syntax measures reveal that smaller minimum room sizes result in lower integration and connectivity but increase interaction hotspots. This study demonstrates the potential of BSP in generating varied game maps and the utility of Space Syntax for assessing their spatial properties
Pengembangan Web Antrian Terapi RSUD Syarifah Ambami Rato Ebu Menggunakan Waterfall dan SUS
No full textDuring the pandemic, the COVID-19 virus spread very quickly through the air so that the government implemented a social distancing policy. However, the high number of patients in the therapy waiting room of Syarifah Ambami Rato Ebu Bangkalan Hospital is feared to make the policy not run optimally. The purpose of this study is to create an online therapy queue system abbreviated as SMART in order to reduce crowds in hospitals. In developing SMART, the Waterfall method is used so that it is sequential starting from system needs analysis, design, to implementation and maintenance. The results of functional testing show that all application features can run well. Furthermore, the usability evaluation using the System Usability Scale (SUS) method produced a score of 70, which indicates that the system has a good level of acceptability and usability. Other values obtained in the Acceptability Ranges are Marginal, the Grades Scale value is C, the Adjective value is Good, and the Promoters and Detractors value is Passive. The implementation of this SMART system has the potential to increase the operational effectiveness of hospitals in managing patient flow and significantly improve user experience through usability evaluation