In the plant factory, plants carry out the process of photosynthesis using artificial light sources in the form of growth lights instead of sunlight. Each type of lamp will emit light with a different wavelength, so the effect on the photosynthesis process is also different. This study examines the pattern of canopy area expansion on plants in response to the light quality of artificial light in the plant factory. Canopy area of Bok choy (Brassica rapa subsp. Chinensis) plant samples for white LED dan red-blue LED treatments were measured non-destructively every two and three days using the Easy Leaf Area application. The growth data is then modeled using linear and polynomial regression. The results showed that the canopy area of the plant in the plant factory increased with the growth period. The plants exposed to red-blue LED lights grew better than plants exposed to white LED, indicating that red-blue lights supported the photosynthesis process better than white light. The regression results also show that the growth of plants in plant factories with a combination of red-blue LED lights is better in terms of MAPE (Mean Absolute Percentage Error) values. The model validation and prediction are also better using red-blue LED than white LED. Based on prediction using the best regression model, the result shows that red-blue LED as artificial light sources better support plant growth in plant factories than white LED lamps.

Several factors influence plant growth, including sun intensity, nutrient content, soil moisture, temperature, genes, and hormones. Many studies have been carried out in constructing plant growth models to simulate plant growth in different treatments. This study aims to develop a mathematical model with a linear regression approach and an artificial neural network. This research method used an experimental design using three treatments consisting of control (T1), 50% shade (T2), and 80% shade (T3). Each treatment had five replications of the chili plant. The tools and materials used were red chili (Capsicum annuum L.) seeds of 30 DAP, a greenhouse of 3 x 3 meters, a drip irrigation control system, 25 x 30 cm polybags, and fertile soil media. The results showed that linear regression models of the 1 st and 2 nd order could be used to predict plant growth with an average RMSE value of 1.53. In contrast, the use of artificial neural networks showed a smaller RMSE value of 0.12 which means that the artificial neural network method was better at predicting plant growth.