Adapun pengertian dari UMKM (Usaha Mikro Kecil Dan Menengah) adalah usaha produktif yang dimiliki oleh perorangan maupun badan usaha yang telah memenuhi kriteria yang ditetapkan agar dapat disebut sebagai usaha mikro. Secara lebih jelas, pengertian UMKM dan kriteria usaha yang termasuk didalam UMKM diatur dalam Undang-Undang Republik Indonesia No. 20 Tahun 2008 tentang UMKM. Dalam undang-undang tersebut disebutkan bahwa UMKM diklasifikasikan berdasarkan jenis usahanya yakni usaha mikro, usaha kecil dan usaha menengah

Sentiment classification plays a crucial role in evaluating user feedback. Today, online media users can freely provide their reviews with few restrictions. User reviews on social media are often disorganized and challenging to classify as positive or negative comments. This task becomes even more difficult when dealing with large amounts of data, making sentiment classification necessary. Automating sentiment classification involves text classification processes, commonly performed using deep learning methods. The classification process using deep learning models is closely tied to text representation. This step is critical as it affects the quality of the data being processed by the deep learning model. Traditional text representation methods often overlook the contextual meaning of sentences, leading to potential misclassification by the model. In this study, we propose a novel fusion text representation model, GloWord_biGRU, designed to enhance the contextual understanding of sentences for sentiment classification. Firstly, we combine the advantages of GloVe and Word2Vec to obtain richer and more meaningful word representations. GloVe provides word representations based on global frequency statistics within a large corpus, while Word2Vec generates word vectors that capture local contextual relationships. By integrating these two approaches, we enhance the quality of word representations used in our model. During the classification stage, we employ biGRU, considering the use of fewer parameters, which consequently reduces computational requirements. We evaluate the proposed model using the IMDB dataset. Several scenarios demonstrate that our proposed model achieves superior performance, with an F1 score of 90.21%.

Japanese language has complex writing systems, Kanji and Kana (Katakana and Hiragana). Each one has different style of writing. One simple way to differentiate is Kanji have more strokes than Kana. Meanwhile, it needs a lot of effort to remember characters of Katakana and Hiragana, thus it will be very difficult to distinguish handwritten Katakana and Hiragana, since there are a lot of similar characters. This is the reason why we need pattern recognition. Handwriting recognition has been the object of intensive research during the last decades, including Japanese characters recognition. In this research, we build an application to recognize Kana's character handwriting. Pattern recognition on Kana is started based on Optical Character Recognition (OCR) consist of scan image, preprocessing, feature extraction and post-processing. Classification process will be done in post-processing using neural network back propagation. The processing of input images, involves training and testing image. Training image uses 3.956 characters done in three phases of training. Testing image uses 552 characters of some images inputted in training phase and image which has not been used before. The test result achieve maximum accuracy of recognition system at 91.88% from train image and 31.03% from test image.

Early detection of suicide risk plays a vital role in improving mental health care outcomes, especially for individuals suffering from depression. Traditional methods, such as clinical interviews, are widely regarded as the gold standard in suicide risk assessment. However, these methods often rely heavily on verbal communication and subjective interpretation, which may lead to the overlooking of important nonverbal cues that could indicate distress. To address this limitation, this study proposes a novel approach to early suicide risk detection by utilizing thermal facial imaging combined with a deep learning model called CapsuleThermNet. CapsuleThermNet is designed by integrating CNN with Capsule Networks (CapsNet). In this architecture, CNNs are responsible for extracting spatial features from thermal facial images, while CapsNet captures hierarchical relationships and performs classification tasks. The study uses a specially developed primary dataset, known as the FaceTherm Suicide Risk (FTSR) Dataset, which includes 152 thermal images of depressed patients. These images are categorized equally into two groups: patients identified as at risk of suicide (76 images) and those not at risk (76 images). The proposed model achieves excellent performance with an accuracy of 99.67%, precision of 93.75%, sensitivity of 100%, specificity of 93.33%, F1-score of 96.77%, and AUC of 0.97. The perfect recall rate highlights the model’s strength in minimizing false negatives—crucial in clinical settings. With its robust classification capabilities, CapsuleThermNet holds promise as a reliable decision-support tool for psychiatrists, enabling earlier detection and timely intervention for patients at risk of suicide before proceeding to more extensive clinical evaluation.

The Point of Sales application, which only focuses on Business to Customer services, has been widely used. On the other hand, it is still rare to find applications that perform Business to Business, which allows users to perform self-management, thus providing convenience to the corporation in storing the data used and utilizing the organization's data sources, as well as being able to integrate non-cash payments needed to increase the number of transactions. This article presents the design of the cashierin point of sales application, which integrates the payment gateway function in the KasirIn point of sales application. Mobile application-based system design, using the scrum development method. The system modeling uses the Unified modeling language (UML), with the interface programming language using Dart and the Flutter Framework. The backend system uses python flask which functions as an API and is connected to the midtrans payment gateway API, with the PostgreSQL database. The system model developed can make it easier to make sales transactions using digital payment methods. Development using the scrum method is also very effective for the software development life circle, because the work management is calculated based on the weight of points from the development team.Keywords: Cashier application, Payment gateway, Scrum, Unified modeling language (UML). ABSTRAK. Aplikasi Point of Sales yang hanya berfokus kepada pelayanan Business to Customer, telah banyak digunakan. Sebaliknya, masih jarang ditemukan apllikasi yang melakukan Business to Business, yang memberikan pengguna melakukan self management, sehingga memberikan aspek kenyamanan kepada koorporasi dalam menyimpan data yang digunakan dan memanfaatkan sumber data organisasinya, serta dapat melakukan integrasi pembayaran secara non tunai yang diperlukan dalam meningkatkan jumlah transaksi. Artikel ini menyajikan rancangan aplikasi point of sales kasirin, yang mengintegrasikan fungsi payment gateway pada aplikasi point of sales KasirIn. Perancangan sistem berbasis aplikasi mobile, dengan menggunakan metode pengembangan scrum. Pemodelan sistem menggunakan Unified modelling language (UML), dengan bahasa pemrograman antarmuka menggunakan Dart dan Framework Flutter. Adapun backend system menggunakan python flask yang berfungsi sebagai API dan terhubung dengan API payment gateway midtrans, dengan database PostgreSQL. Model sistem yang dikembangkan ini dapat mempermudah melakukan transaksi penjualan dengan metode pembayaran secara digital. Pengembangan dengan metode scrum juga sangat efektif untuk software development life circle, karena manajemen pekerjaannya dihitung berdasarkan bobot poin dari tim pengembang.Kata Kunci: Aplikasi kasir, Gerbang pembayaran, Scrum, Unified modeling language (UML).

The growing sophistication of generative Artificial Intelligence (AI) has intensified the threat posed by deepfake technologies, which are capable of producing highly realistic yet fabricated facial images and videos. These manipulated visuals can mislead the public, infringe on personal privacy, and damage reputations. This study aims to develop an effective deepfake image detection system using Convolutional Neural Networks (CNN) enhanced with EfficientNet architectures (B3–B5). The research adopts the Cross Industry Standard Process for Data Mining (CRISP-DM) methodology, providing a structured data science framework that spans from problem definition to deployment. Three open-access datasets (Celeb-DF v2, DeeperForensics-1.0, and DFDC) are utilized to train and evaluate the models. Experimental results show that EfficientNet-B5 achieves the highest classification accuracy at 93.2%, outperforming both the baseline CNN and other EfficientNet variants. The proposed method demonstrates strong cross-dataset generalization and computational efficiency, making it suitable for real-world applications. This research contributes a comparative evaluation of scalable deepfake detection models, practical deployment insights, and a foundation for future work in explainable and real-time AI-based media forensics.