Claim Missing Document
Check
Articles

Found 3 Documents
Search
Journal : Jurnal Teliska

Aplikasi Sensor Ultrasonik Dalam Pembacaan Level Air Pada Sistem Pertanian Aquaponic Daniesar, Muhammad Nouval; Dewi, Tresna; Oktarina, Yurni
JURNAL TELISKA Vol 16 No I (2023): TELISKA Maret 2023
Publisher : Teknik Elektro Polsri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.5281/zenodo.8031245

Abstract

IMPLEMENTASI ALGORITMA POLYNOMIAL REGRESSION UNTUK PREDIKSI PERTUMBUHAN TANAMAN PAKCOY PADA SISTEM HIDROPONIK Yurni Oktarina; Rendi Dwi Yanto; Tresna Dewi
JURNAL TELISKA Vol 19 No I (2026): TELISKA Maret 2026
Publisher : Teknik Elektro Polsri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36257/teliska.v19iI.11816

Abstract

This study aims to apply the polynomial regression algorithm to model and predict the growth of pakcoy plants in a hydroponic system. The observed growth parameters include plant height, plant width, and number of leaves, with plant age used as the independent variable. Data were collected over two planting periods, with weekly observations conducted from the seedling stage until harvest. In addition to morphological parameters, Total Dissolved Solids (TDS) and water temperature were recorded as supporting parameters to ensure stable cultivation conditions throughout the study. The non-linear relationship between growth parameters and plant age was represented using a second-order polynomial regression model. The modeling results indicate a good level of fit, with coefficients of determination (R²) of 0.989 for plant height, 0.946 for plant width, and 0.970 for number of leaves, respectively. The relatively low Root Mean Square Error (RMSE) values for each parameter indicate that the model is capable of providing predictions with low estimation error. These findings demonstrate that second-order polynomial regression is a simple and effective approach for modeling the growth dynamics of pakcoy plants in hydroponic systems with limited data availability
PREDIKSI PEMANENAN ENERGI PADA ARRAY PIEZOELEKTRIK KONFIGURASI SERI-PARALEL BERBASIS RANDOM FOREST REGRESSION Fairuz Attalah; Yurni Oktarina; Pola Risma; Assyifa Mourlina Faraquinsha; Hendra Marta Yudha
JURNAL TELISKA Vol 19 No I (2026): TELISKA Maret 2026
Publisher : Teknik Elektro Polsri

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36257/teliska.v19iI.11829

Abstract

This study analyses the electrical characteristics and predicts the power output of a piezoelectric smart carpet system through a comparative analysis of series and parallel circuit configurations using the Random Forest Regression (RFR) algorithm. The system was designed using 16 piezoelectric ceramic elements per block integrated into a 100 cm carpet structure. Experimental data were collected from five subjects with body masses ranging from 54–98 kg through walking and running activities, yielding 100 observational samples. Results show the series configuration produced an average power output of 1649.3 mW, outperforming the parallel configuration by 104.8%, which yielded only 805.1 mW, with peak voltage reaching 80 V during running. The RFR model optimized using GridSearchCV with 5-fold cross-validation achieved a coefficient of determination (R²) of 88.25%, a Mean Absolute Error (MAE) of 145.06 mW, and a Root Mean Squared Error (RMSE) of 179.15 mW. Feature importance analysis revealed that circuit configuration (47.59%) and step frequency (47.18%) are the dominant predictive factors, while body mass contributed only 5.22% due to mechanical saturation in the carpet structure. This study confirms that RFR is effective as a predictive model for optimizing biomechanical energy harvesting systems in public infrastructure.