Jurnal Informatika: Jurnal Pengembangan IT
Not a member yet
437 research outputs found
Sort by
Pengembangan Aplikasi Chatbot Untuk Layanan Penerimaan Mahasiswa Baru Berbasis Natural Language Processing
No full textAbstrak – Penelitian ini bertujuan untuk mengembangkan dan mengimplementasikan sistem chatbot berbasis Natural Language Processing (NLP) untuk mendukung layanan informasi Penerimaan Mahasiswa Baru (PMB) di Universitas Janabadra. Layanan PMB selama ini masih bergantung pada interaksi manual yang terbatas pada jam kerja. Oleh karena itu, diperlukan solusi digital yang mampu memberikan informasi secara cepat, akurat, dan real-time. Sistem dikembangkan menggunakan framework CodeIgniter 4 dan memanfaatkan algoritma Naive Bayes untuk klasifikasi intent serta Levenshtein Distance untuk pencocokan kemiripan teks. Dataset pelatihan disusun berdasarkan kumpulan pertanyaan umum calon mahasiswa. Hasil evaluasi menunjukkan bahwa chatbot mampu menjawab 70% dari 500 pertanyaan secara otomatis dengan akurasi 92% dan waktu respons rata-rata 0,5 detik. Selain itu, chatbot mampu menurunkan beban kerja staf administrasi hingga 30%. Survei terhadap 100 pengguna menunjukkan bahwa 85% responden merasa puas terhadap kecepatan dan keakuratan respons sistem. Sistem ini juga mendukung penyimpanan konteks percakapan dan integrasi langsung dengan informasi PMB universitas. Penelitian ini menyimpulkan bahwa chatbot berbasis NLP dapat menjadi solusi efektif dalam meningkatkan efisiensi layanan informasi pendidikan tinggi. Pengembangan lanjutan diarahkan pada perluasan dataset, adopsi model NLP berbasis Transformer, serta integrasi lintas platform komunikasi untuk memperluas jangkauan layanan. Kata Kunci: Chatbot, Natural Language Processing, Naive Bayes, Levenshtein Distance, Penerimaan Mahasiswa Baru
Perbandingan Logistic Regression, SVM, dan Random Forest untuk Analisis Sentimen Ulasan Aplikasi Gopay
No full textThe expansion of Indonesia's digital financial landscape has triggered a surge in the adoption of e-wallets, most notably GoPay. Within this context, feedback available on application platforms such as the Google Play Store serves as a crucial metric for assessing user sentiment and service quality. Sentiment analysis based on machine learning algorithms allows for systematic and objective identification of public opinion. This study used 3,000 user reviews collected through web scraping from the Google Play Store, received up to April 21, 2025, with initial labeling based on a lexicon approach. Although many studies have compared sentiment classification algorithms, there has been no research specifically comparing the performance of Logistic Regression, Support Vector Machine (SVM), and Random Forest in the context of GoPay user reviews with lexicon based labeling. This paper aims to fill the existing void by evaluating the comparative performance of three algorithms based on sentiment classification metrics. Preprocessing procedures encompassed cleaning, case-folding, stemming, slang normalization, tokenizing, filtering, and labeling to ensure data quality. The models, built within the Scikit-learn environment, were tested for accuracy, precision, recall, and F1-score. Empirical results confirm that Logistic Regression outperformed the alternatives, securing 88.16% accuracy while maintaining stability across all sentiment categories. SVM recorded 87.5% accuracy but was weak in detecting negative sentiment. Random Forest showed the lowest performance with 79.33% accuracy and less consistent classification results. Thus, Logistic Regression is recommended as the most effective algorithm for GoPay user sentiment analysis. Future research can explore deep learning-based approaches to handle higher sentiment complexity
Segmentasi Pelanggan Berdasarkan Model LRFM Menggunakan Algoritma K-Means dan Optimasi Klaster Dinamis
No full textThe number of tax training participants often does not meet the minimum quota, resulting in the cancellation of many training classes. Throughout 2022, there were 27 training classes that failed to take place due to a lack of participants. One of the reasons is that promotions have not utilised historical customer data to set marketing targets more precisely. By utilising historical customer data, companies can design more targeted promotional strategies and increase the number of training participants. Therefore, this research aims to segment customers using the dynamic K-Means algorithm based on the Length, Recency, Frequency, and Monetary (LRFM) model, so that customer behaviour patterns when registering for training can be identified. The clustering results are then visualised to facilitate analysis and decision-making. This research resulted in three customer segments, namely Loyal customers (Gold, 17%), Lost customers (Diamond, 64%), and New customers (Silver, 17%). With this segmentation, it is expected that the company can conduct more effective promotions and increase the number of trainees in the future
Implementasi Naïve Bayes untuk Rekomendasi Pembelian Produk pada Aplikasi E-commerce
No full textElectronic commerce (e-commerce) is a platform that influences buying and selling habits in Indonesia, with data from the Central Statistics Agency 2023 showing 31,753 e-commerce businesses using consumer review data as a determinant of product and service quality. This research aims to develop a sentiment-based product recommendation system using the Naïve Bayes algorithm. The research methodology includes collecting 1,287 data samples obtained from customer reviews using Web Scraper technology on the official MSI Official Store e-commerce platforms in the Tokopedia, Shopee, and Blibli applications. The results of data preprocessing yielded 921 clean data, and the Naïve Bayes Algorithm was applied as a classification model and system implementation in a website application. The data was then divided into 80% for training and 20% for testing. Model evaluation showed an accuracy of 82% for training data and 71% for testing data. These results indicate the effectiveness of the Naïve Bayes algorithm in forming a sentiment-based product recommendation system. This recommendation system helps users make more informed purchasing decisions based on consumer sentiment analysis. This research contributes to the development of intelligent recommendation systems that can improve user decision-making in the digital marke
Transformasi Digital Sistem Kehadiran untuk Budaya Hybrid Work dengan TOGAF Framework
No full textSince the end of the COVID-19 pandemic in 2023, hybrid work or work from anywhere (WFA) methods have become increasingly popular. While this work culture offers flexibility, it also poses challenges for companies in monitoring employee attendance. This study proposes the implementation of the TOGAF Framework to design an attendance management system that supports a hybrid work culture. The design process begins with a preliminary phase to identify user needs and the limitations of traditional systems. The system is designed using TOGAF ADM (Architecture Development Method), covering phases such as Architecture Vision, Business Architecture, Information System Architecture, and Technology Architecture. Technologies such as IoT, GPS, facial recognition, and mobile applications are employed to ensure system flexibility and accuracy. Testing is conducted in two stages: Blackbox Testing to verify functionality against specifications, and User Acceptance Testing (UAT) to evaluate system usability in real-world conditions. The test results show that the system meets all specifications, improves operational efficiency by up to 40%, and ensures the security and accuracy of attendance data. The system is also designed for integration with other modules, such as payroll and HRIS, to support strategic decision-making. This approach provides an effective and adaptive solution for monitoring employee attendance, whether working remotely or in the office, and enhances productivity in modern work environments
Pengembangan Aplikasi Presensi QR Code Berbasis Website Dengan Metode Agile
No full textAttendance recording for students at Pondok Pesantren Mahasiswa (PPM) Al-Hikmah Semarang is still conducted manually, making it prone to recording errors and time-consuming when compiling attendance data. This study aims to develop a QR Code-based attendance system to improve the efficiency and accuracy of attendance recording. The method used involves designing and developing a web-based system using the Laravel framework and the Agile methodology. The system is designed to be used by both students and class supervisors during learning activities at the dormitory. The research results show that the system can automate student attendance through QR Code scanning, store data in a structured manner, and provide accurate attendance reports that are easily accessible to the dormitory administrators. Additionally, features such as schedule management, student data management, and attendance reporting based on specific criteria have been implemented to support more effective administration. The system is also equipped with a feature to print attendance recap reports. With the implementation of this system, student attendance management becomes more efficient, transparent, and less prone to errors compared to the previously used manual method
Verifikasi Wajah untuk Menghitung Jumlah Transaksi Pengunjung Menggunakan Metode Deep Metric Learning
No full textThis research carries the theme of facial recognition to detect visitors' faces by counting the number of times visitors make transactions. The objective of this research is to develop and implement a face verification system for public purposes, such as commercial purposes. One potential application of this system is in the realm of promotions, where it could be utilized to track the number of transactions conducted by visitors. The method employed utilizes deep metric learning (DML) to generate a model capable of verifying various facial images through the Convolutional Neural Network (CNN) architecture, which is designed to train human face image data. The triplet loss method is employed in training data due to its recognition as a more flexible approach in utilizing labels (in the form of face images) to facilitate comparison with the detected face images. The model employed for face recognition applications is facenet, a system that has been demonstrated to achieve a high degree of accuracy. The research's output is an application capable of swiftly and precisely verifying facial images of visitors and calculating the number of visitor transactions. The number of visitor transactions can subsequently be utilized as a promotional or discount strategy in commercial services
Integrasi Backend Golang-Echo pada Aplikasi Greenly sebagai Solusi Teknologi Pengelolaan Sampah Digital
No full textThe waste problem in Indonesia is increasingly pressing, with 35.7% of the 31.9 million tons of national waste in 2023 not being managed properly. This study develops the Greenly web application backend as a digital solution to support waste reporting, recycling education, and increasing community participation through a gamification system. The methodology used is the Waterfall model, including needs analysis, design with Entity Relationship Diagram (ERD), implementation, and testing. The backend is built using the Golang and Echo frameworks, then packaged in Docker and deployed on the AWS EC2 service. The Continuous Integration/Deployment (CI/CD) process is carried out using GitHub Actions, with Nginx as a reverse proxy. Testing is carried out through Integration Test to ensure the reliability of key features such as CRUD data, waste reporting, and gamification. The results show that the backend system runs stably, safely, and efficiently, with an automatic CI/CD flow that is successfully executed without errors. The main contribution of this study is the provision of an adaptive and reliable backend as the foundation for a digital waste management system based on community participation
Perancangan Visualisasi Elektronik Pencegahan Jantung Koroner Berbasis Teknologi Augmented Reality
No full textAugmented reality (AR) merupakan teknologi yang memungkinkan benda maya yang dimasukan secara real time di dalam dunia nyata. Teknologi augmented reality sangat menarik dan mudah untuk diterapkan dan digunakan bagi penggunanya. Pemanfaatan teknologi augmented reality pada saat ini telah meluas ke berbagai aspek termasuk kesehatan. Di dunia kesehatan, penyakit jantung koroner masih menjadi ancaman di Indonesia bahkan di dunia. Penyakit jantung koroner adalah gangguan fungsi jantung, akibat otot jantung kekurangan nutrisi dan oksigen karena adanya penyempitan pada pembuluh darah koroner, namun masyarakat masih sering menilai bahwa penyakit jantung merupakan penyakit yang dapat sembuh dengan sekali pengobatan. Maka di butuhkan suatu media informasi yang kuat, yang dapat meningkatkan pengetahuan masyarakat mengenai penyakit jantung khususnya penyakit jantung koroner karena penyakit jantung jenis ini yang paling populer di masyarakat. Penelitian ini bertujuan untuk mengembangkan sebuah aplikasi berbasis augmented reality (AR) pada platform Android yang dapat digunakan sebagai media visualisasi penyakit jantung, khususnya penyakit jantung coroner. Hal ini untuk meningkatkan pengetahuan masyarakat mengenai penyakit jantung khususnya penyakit jantung coroner. Dalam membangun aplikasi ini, metode yang digunakan adalah metode Multimedia Development Life Cycle (MDLC). Dari hasil pengujian yang dilakukan dengan metode Alpha dan Beta, dapat disimpulkan bahwa aplikasi ini dinilai layak untuk digunakan sebagai media edukasi interaktif tentang penyakit jantung koroner, hal ini dibuktikan dengan hasil pengujian beta test yang telah dilakukan dan menghasilkan nilai rata-rata tingkat persetujuan usability aplikasi secara keseluruhan sebesar 89,3 %
Implementasi Algoritma Support Vector Regression untuk Prediksi Harga Emas Berdasarkan Data Historis
No full textAmidst global economic volatility, accurate forecasting of gold prices remains a crucial and challenging task for investors and financial policymakers, as gold functions as a vital safe-haven asset and a hedge against inflation. This study focuses on gold price prediction utilizing the Support Vector Regression (SVR) algorithm, with the main objective of improving forecast accuracy. The relevance of this prediction is underpinned by the dynamic characteristics of gold prices, which is essential for decision-making by various stakeholders. Historical gold price data were obtained from the investing.com platform. The SVR implementation was carried out utilizing the Radial Basis Function (RBF) kernel. The SVR parameter optimization process employing Grid Search successfully identified the optimal values, namely C=1000, ϵ=0.5, and γ=0.01. To ensure model robustness and generalization capability, validation was performed using 5-Fold Cross Validation, which yielded an average Mean Absolute Percentage Error (MAPE) of 0.66%. The very high level of SVR accuracy, alongside its consistency across each fold, stability, and reliability, indicates that the optimized SVR model is a prospective solution for gold price forecasting in the commodity market