Predicting Optimum Air Quality for Lactuca sativa L. var. Longifolia (Romaine Lettuce) Growth in a Vertical Hydroponics Farm: A Machine Learning Approach

Proponent/s Name/s (Last Name, First Name, Middle Initial)

• Angela Charlize D. Calingo, De La Salle University, ManilaFollow
• Danielle Jashean T. Cunanan, De La Salle University, ManilaFollow
• Rhianne Daphne O. Sia, De La Salle University, ManilaFollow
• Calianne Lucille P. Sy, De La Salle University, ManilaFollow
• Aaron Don M. Africa, De La Salle University, ManilaFollow

Document Types

Paper Presentation

Research Theme (for Paper Presentation and Poster Presentation submissions only)

Sustainability, Environment, and Energy (SEE)

School Name

De La Salle University, Manila

Track or Strand

Science, Technology, Engineering, and Mathematics (STEM)

Research Advisor (Last Name, First Name, Middle Initial)

Africa, Aaron Don, M.

Start Date

23-6-2026 1:30 PM

End Date

23-6-2026 3:00 PM

Zoom Link/ Room Assignment

DLSU Manila Campus (In-person) - Brother Andrew Gonzalez Multipurpose Hall, 20th floor

Abstract/Executive Summary

Controlled Environment Agriculture (CEA) uses technology to regulate growing conditions to improve crop yield. Hydroponics, a type of CEA, is a soilless plant-growing method that suspends plants and delivers a nutrient-rich solution to the roots. While most studies on hydroponics focus on optimizing water and nutrient delivery, few examine the effects of air quality on plant growth. Key components, such as carbon dioxide (CO₂), total volatile organic compounds (TVOCs), particulate matter (PM2.5 and PM10), humidity, and temperature, are often excluded from machine learning (ML) models that predict crop outcomes. This study aims to address this by training ML models using air quality data collected in a residential condominium in Metro Manila. Real-time sensors monitored conditions in a vertical hydroponics system with an open-type growth chamber for plant exposure to urban air conditions. Lactuca sativa L. var. Longifolia was planted and monitored over a 3-week period.  Air quality variables and plant morphological properties were recorded at fixed intervals. The model, eXtreme Gradient Boosting (XGBoost), was used for predictive analysis and evaluated using Coefficient of Determination (R²), Root Mean Squared Error (RMSE), and Mean Absolute Error (MAE).  Results showed that optimal air quality ranges include: 40-63µg/m3 PM2.5, 39-71µg/m3 PM10, 403-419ppm CO₂, 0.008 - 0.012 mg/m3 TVOC, 25-30°C temperature, and 55-76% humidity. Moreover, using morphological and air quality data as inputs to the ML models, this research concludes that the XGBoost model achieved the highest predictive accuracy, demonstrating the most balanced performance.

Keywords

air quality; machine learning; XGBoost; hydroponics; urban farming

Initial Consent for Publication

yes

Statement of Originality

yes