19 research outputs found
Hybrid Deep Learning Models Using LSTM with Random Forest for Radio Frequency-Based Human Activity Recognition in Line-of-Sight and Non-Line-of-Sight Environments
No full textHuman Activity Recognition (HAR) has become an important field of study because of its wide range of applications in healthcare, security, and smart living systems. Radio Frequency (RF)-based HAR offers a non-invasive and privacy-preserving alternative to traditional vision-based systems. This study proposes a hybrid deep learning model combining Long Short-Term Memory (LSTM) networks with Random Forest classifiers for RF-based HAR, aiming to improve recognition accuracy across diverse environments. The model was evaluated using Channel State Information (CSI) and Received Signal Strength Indicator (RSSI) features under Line-of-Sight (LOS) and Non-Line-of-Sight (NLOS) conditions. Synthetic Minority Over-sampling Technique (SMOTE) was integrated to balance the dataset, and K-fold Cross-Validation was employed to assess robustness. The dataset included data from 8 subjects performing 10 different activities. The model achieved high classification accuracy, with 99.40% in Environment 1 (LOS), 97.58% in Environment 2 (LOS), and 98.30% in Environment 3 (NLOS), demonstrating the models adaptability and effectiveness. The results highlight the potential of the hybrid LSTM with Random Forest approach for scalable and reliable RF-based HAR systems that can be integrated into real-world Internet of Things (IoT) applications
Soybean Collect Recommender Based on Distance and Productivity Cluster Using K-means Clustering and Simple Addictive Weighting Method
No full textSoybeans are an essential agricultural product that is one of the primary food sources in Indonesia, such as tempeh, tofu, soy milk, soy sauce, and other preparations. However, production yields, harvested land area, and soybean productivity in each district or city in Central Java Province vary widely. Differences in soybean productivity in each area are due to production factors such as area, use of fertilizers, seeds, and labor. This study tries to provide recommendations for soybean harvesting based on the distance and productivity of an area using K-means clustering and the simple addictive weighting method. In the Central Java Province, 35 regions will be divided into four clusters: the first with high productivity, the second with medium productivity; the third with low productivity; and the fourth with very low productivity. Additionally, based on the fourth cluster clustering results, it will be advised to take soybeans from other clusters by taking the closest distance and cluster members into account. According to the research, four clusters have formed: the first has five members, the second has fourteen, the third has nine, and the fourth has seven. The fourth cluster, which consists of seven members who do not grow soybeans, is advised to buy soybeans from the following regions: Kendal Regency, Klaten Regency, Magelang Regency, Batang Regency, and Brebes Regency
Pemeriksaan Pola Kalimat Otomatis Pada Sebuah Karangan Menggunakan POS Tagging Bahasa Indonesia Dan LALR Parser
No full text Dalam era perkembangan teknologi yang pesat ini, berbahasa mempunyai peran penting dalam kehidupan sehari-hari seperti untuk berkomunikasi dengan sesama secara lisan maupun tulisan. Komunikasi akan berlangsung dengan baik jika bahasa yang digunakan dapat dipahami sehingga pesan dapat tersampaikan. Dalam komunikasi tulisan, keterampilan menulis diperlukan untuk menghasilkan tulisan yang dapat menyampaikan pesan dengan baik. Salah satu bentuk hasil dari keterampilan menulis adalah sebuah karangan. Penulisan karangan harus memperhatikan kaidah pemakaian bahasa yaitu fonologi, morfologi, dan sintaksis. Pentingnya kaidah tersebut khususnya sintaksis atau struktur dan pola kalimat dapat mengungkapkan ide yang dapat tersampaikan dengan baik dan mudah untuk dipahami melalui karangan. Penelitian ini bertujuan untuk membantu dalam memeriksa pola kalimat pada sebuah karangan secara otomatis. Dalam pemeriksaan ini diimplementasikan dengan bahasa pemrograman python pada jupyter notebook menggunakan library nltk untuk proses preprocessing, library flair nlp untuk proses part of speech tagging bahasa Indonesia dan penggunaan tabel lalr parser untuk pemeriksaan pola kalimat. Pola kalimat yang digunakan pada pemeriksaan ini adalah S-P, S-P-O, S-P-K, S-P-O-K, S-P-Pel-K, dan S-P-O-Pel-K. Hasil dari penelitian ini adalah berupa pemeriksaan pola kalimat otomatis pada sebuah karangan sederhana dengan batasan menggunakan kalimat tunggal dan kalimat aktif. Pemeriksaan ini dapat memeriksa 14 dari 16 kalimat pada karangan dengan nilai keberhasilan sebesar 87,5% dan nilai keakuratan sebesar 62,5%. Faktor yang mempengaruhi hasil tersebut adalah variasi komponen pola kalimat yang masih terbatas dan penggunaan flair nlp dalam proses pos tagging yang dapat menghasilkan label jenis yang berbeda pada suatu kata yang dipengaruhi oleh letak posisi kata tersebut pada sebuah kalimat
Comparative Analysis of Hybrid Intelligent Algorithms for Microsleep Detection and Prevention
No full textMicrosleep is a critical factor contributing to traffic accidents, posing significant risks to road safety. Research by the AAA Foundation for Traffic Safety found that 328,000 sleep-related driving accidents happen annually in the United States, underscoring the widespread and dangerous nature of drowsy driving. These incidents often occur without warning, making them especially hazardous and difficult to prevent through conventional means alone. This research aims to improve the accuracy of microsleep detection by developing a hybrid intelligent algorithms. It compares three intelligent algorithms: Fuzzy Logic (FL), representing scheme A; Fuzzy Logic combined with Artificial Neural Networks (FL-ANN), representing scheme B; and a combination of Fuzzy Logic, ANN, and Decision Trees (FL-ANN-DT), representing scheme C. These methods were evaluated using performance metrics such as MSE, MAE, RMSE, R², and response time. The results indicate that Scheme C (FL-ANN-DT) significantly outperforms the other approaches, achieving an MSE of 5.3617e-32, MAE of 4.3823e-17, R² of 1.0, and an RMSE close to zero, demonstrating near-perfect accuracy. Compared to previous models, this hybrid approach enhances prediction precision while maintaining real-time feasibility. The findings highlight the potential of FL-ANN-DT as an advanced microsleep detection system, contributing to improved road safety and real-time monitoring applications. This system can serve as a proactive safety layer in driver assistance technologies, reducing the risk of fatigue-related accidents and potentially saving lives
Prediksi Kabut Menggunakan Recurrent Neural Network dengan Attention Mechanism di Bandara Ruteng
No full textFenomena kabut menjadi tantangan signifikan dalam operasional penerbangan, terutama di wilayah dengan topografi kompleks seperti Bandara Ruteng. Penelitian ini bertujuan untuk mengembangkan model prediksi kabut menggunakan Recurrent Neural Network (RNN) dengan integrasi Attention Mechanism (AM) untuk menangkap pola temporal dalam data cuaca yang kompleks. Data penelitian mencakup 61.187 dengan parameter cuaca yang direkam setiap jam selama sepuluh tahun terakhir. Model dilatih selama 50 epoch, ukuran batch 32 dan optimizer Adam . Hasil pengujian menunjukkan bahwa model RNN+AM memiliki performa lebih baik dibandingkan model RNN, dengan nilai MAE sebesar 0,27% dan RMSE sebesar 5,19%, lebih rendah dibandingkan RNN dengan MAE sebesar 0,44% dan RMSE sebesar 6,64%. Evaluasi confusion matrix menunjukkan bahwa RNN+AM memiliki tingkat kesalahan False Positive dan False Negative yang lebih kecil, dengan akurasi yang lebih tinggi dalam prediksi kabut. Integrasi AM terbukti meningkatkan kemampuan model dalam memprioritaskan fitur relevan yang mendukung akurasi prediksi
