This study compares the performance of two deep learning models, namely Convolutional Long Short-Term Memory (ConvLSTM) and Long-term Recurrent Convolutional Network (LRCN), in the task of recognizing human activity from videos. Human activity recognition is an important field in computer vision with many applications, such as security monitoring, human-computer interaction, and social media-based video analysis. ConvLSTM is a model that combines convolution operations with long-term memory LSTM, thus capable of capturing spatial and temporal information simultaneously. This approach is ideal for processing video data sequences that have spatial and temporal dimensions. On the other hand, LRCN combines the power of spatial feature extraction from Convolutional Neural Network (CNN) and temporal sequence modeling through Recurrent Neural Network (RNN), specifically LSTM, to understand movement patterns in videos. The study used the UCF50 dataset consisting of 50 activity classes, but was limited to five classes for the focus of the experiment. The dataset was divided into 80% for training and 20% for testing, and the model was drilled for 50 epochs using early stopping to prevent overfitting. The results show that both models have high training performance. ConvLSTM achieved a training accuracy of around 98% and a validation accuracy of 90%, while LRCN achieved a training accuracy of 99.5% and a validation accuracy of 88%. Although ConvLSTM demonstrated good stability on the validation data, further testing using TikTok videos as real-world data showed that LRCN had a higher confidence level in recognizing activities, with most predictions achieving confidence scores above 80%. This difference in performance indicates that while ConvLSTM excels in generalizing on training data, LRCN is more robust to real-world data variations.

Land degradation challenges rice productivity in Indonesia, demanding innovative agronomic solutions. This study evaluated the interaction between humic acid-coated biochar rates and plant spacing on the growth and yield of rice (var. IR-64). A field experiment using a factorial Randomized Complete Block Design was conducted in Malang. The treatments used a 2 factorial, the first factor included four biochar rates (0, 5, 10, and 15 t ha⁻¹) and the second factor used three plant spacings (20x20, 20x30, and 20x40 cm). The result showed a significant interaction (P 0.05) between biochar rate and spacing for the number of tillers and stem diameter. The highest tiller count (35.27) was recorded in the 5 t ha⁻¹ biochar rate, combined with 20x40 cm spacing. While the main effect of biochar on yield was not significant, plant spacing had a dominant impact; the 20x30 cm spacing yielded the highest grain weight per plot (8.74 kg). These findings highlight that rice yield optimization depends on the specific synergy between soil ameliorants and planting density

This community service aims to develop superior agribusiness products based on local wisdom in Batuputih Laok Village, Batuputih Subdistrict, Sumenep Regency. The implementation method includes five main stages: needs identification and analysis, program and action plan development, training and counseling, program implementation, and monitoring, evaluation, and follow-up. The results show a significant increase in agricultural productivity and quality due to the integration of modern agricultural technology with local knowledge. Corn productivity increased by 30%, and the quality of tobacco produced improved, leading to higher selling prices. Support from the government, research institutions, the private sector, and active community participation were key factors in the program's success. In conclusion, this program successfully improved farmers' welfare and strengthened the village's agribusiness capacity sustainably.

Upaya melakukan analisis emosi pada teks komentar mahasiswa dalam evaluasi pembelajaran sangat penting dilakukan. Komentar dalam kuesener umumnya tidak diolah, padahal  data tersebut mengandung informasi dalam mengungkap emosi mahasiswa dalam proses pembelajaran.  Untuk itu deteksi dan klasifikasi emosi pada opini mahasiswa  dapat memperbaiki hasil kuesioner. Penelitian ini bertujuan menerapkan metode klasifikasi emosi pada teks komentar mahasiswa berbasis pada leksikon emosi  dari NRC Emolex.  Jenis emosi yang akan dideteksi adalah 8 jenis emosi, yaitu marah (Anger), antisipasi (anticipation), jijik (disgust), takut (fear), bahagia (joy), sedih (sadness), terkejut (surprise)   dan yakin (trust) . Data diambil dari komentar dan saran mahasiswa pada kuesioner pada Universitas  AKPRIND Indonesia tahun 2020-2022  sebanyak 4000 data yang telah dilabeli secara manual.   Tujuan lain dari studi ini adalah melihat sejauh mana efektivitas leksikon emosi Emolex untuk klasifikasi emosi teks kuesioner akademis. Hasil penelitian menunjukkan rata-rata akurasi sebesar 56,2%. Dari yang diketahui label emosinya 3 prosentase tertinggi ada pada label Sadness (19,2%), Joy(16,7%) dan Fear (13,5%) yang masing-masing memiliki akurasi 72%, 68% dan 68%. Dari penelitian terungkap bahwa kinerja Emolex untuk klasifikasi emosi masih kurang memuaskan dan memerlukan pengembangan leksikon lebih jauh lagi.   

The aim of this study the first is to know the different effects of treatment with red-blue LED and then compare it to treatment with white fluorescent light for green mustard growth, and the second is for determined the composition from each different treatment to reach the optimum growth of green mustard in the plant factory environment. In this case, treatment with fluorescent light will be used as a control variable for the observation. The research parameter is air temperature, light intensity, and the plant's physical measurements such as diameter and height. This study used a Single Plant Design, the treatment specified in this research was a planting model consisting of the first indoor hydroponic planting model or Plant Factory by applying LED (M1) irradiation, the second outdoor planting model in a greenhouse using full sun irradiation. (M2), and the third cropping model uses direct planting in the field (M3), and the fourth cropping model uses hydroponic planting in open spaces (M4). The result showed that cultivation with the plant factory model produced the best stover wet weight, which was 165.33 g, in contrast to the greenhouse planting model which produced a wet stover weight of 97.33 g and was different from the conventional model, which was 47.33 g. Vegetable wet weight is decent. Consumption also produced the highest with the plant factory planting model, which was 159.67 g per plant, followed by the greenhouse planting model, which was 92.33 g, and different from the conventional model, which was 46.67 g. Whereas if the method converted per ha, the best yields found in the plant factory cultivation model produce caisim vegetables of 27.56 tons/ha, different from the greenhouse planting model is 16.22 tons ha and significantly different from the conventional model which is 7.89 tons/ha.