4 research outputs found
Efficient Waste Classification in Cisadane River Using Vision Transformer and Swin Transformer Architectures
No full textThe increasing volume of waste in rivers has become a serious environmental problem. This study proposes the implementation of Artificial Intelligence (AI)-based models, specifically Vision Transformer (ViT) and Swin Transformer, for an automatic waste sorting system in the Cisadane River, Tangerang. The dataset used combines public sources and field data, processed through preprocessing and augmentation to improve robustness. Model training was conducted using k-fold cross-validation, pruning, and deployment testing on edge devices to ensure generalization and efficiency. Several architectural innovations were introduced, including Dynamic Patch Size for adapting to various waste shapes and sizes, and Spatial-Aware Attention to enhance focus on waste objects against complex river backgrounds. The evaluation involved a confusion matrix and statistical analysis using a paired t-test to validate the significance of the results. Experimental findings show that Swin Transformer achieved the highest accuracy of 94.2%, surpassing ViT at 91.8%, with precision of 93.5%, recall of 92.7%, and F1-score of 93.1%. Swin Transformer also proved more reliable in dynamic lighting and cluttered environments. This study demonstrates the potential of Transformer-based architectures in automatic waste classification, contributing to smarter and more efficient AI-based environmental management technologies
Penentuan Hoax pada Artikel Politik Berbahasa Indonesia di Sosial Media dengan Similarity Jaccard dan Algoritma Stemming
No full textPesatnya perkembangan teknologi di era globalisasi saat ini membawa pengaruh besar dalam proses pencarian informasi. Kita dapat melihat pertumbuhan yang sangat besar pada volume berita online yang tersedia pada jaringan internet, maupun pada jaringan lainnya. Salah satunya adalah sosial media yang memilik banyak informasi yang mengenai artikel-artikel berita atau artikel-artikel informasi lainnya. Berita merupakan sumber informasi mengenai kejadian terkini yang mana dapat ditemukan pada internet dan media sosial. Saat ini berita-berita yang disebarkan terutama berita mengenai politik yang dapat mengakibatkan salah penafsiran karena berita tersebut belum tentu benar atau salah sehingga dibutuhkan pengklasifikasian artikel politik apakah termasuk dalam kategori hoax atau non hoax. Hoax adalah berita kebohongan yang disebar untuk memperoleh kepercayaan agar masyarakat akan merasa yakin bahwa konten tersebut benar. Dampak lain dari hoax dapat merugikan emosi hingga finansial masyarakat. Proses klasifikasi hoax menggunakan tahap preprossessing yang terdiri dari tokenization dan stemming. Dilanjutkan dengan proses pembobotan kata dan jaccard similarity hingga proses klasifkasi dengan menggunakan metode Vector Space Model (VSM). Hasil evaluasi pada penelitian ini menggunakan confusion matrix, dimana diperoleh hasil precision sebesar 0,92 recall sebesar 0,80 dan akurasi didapatkan sebesar 87 %
Prediction Of Students Academic Success Using Case Based Reasoning
No full textAcademic success for a student is influenced by many factors during their study period. Factors such as student gender, student absenteeism, parental satisfaction with schools, relations and parents who are responsible for students can influence student success in the academic field. Researchers try to find out what are the most dominant factors in determining academic success for a student at different levels of education such as elementary, middle and high school level. Previous research grouped the level of student academic success into three levels, namely low, medium, high and obtained 15 Association Rules Generated By Apriori Algorithm. This study tried to find out and predict the possible level of academic success of students by using 9 Association Rules Generated By Apriori Algorithm from previous research. The method used to predict the level of student academic success is case based reasoning with the nearest neighbor algorithm. By using the Association Rules Generated By Image Algorithm and with the data set from the xAPIEducational Mining Dataset the case similarity value was obtained with knowledge data that is 1 with a percentage of 81%, and data that had a similarity value of less than 1 was 19%. While in the previous study the best classification accuracy was 80.6% by the Voting classifier. And the grouping of success data is divided into two, namely low and high
