Analysis of Key Factors and Development of Yield Prediction Model for Greenhouse Cucumber Cultivation

Purpose: The purpose of this study is to examine the relationship between yield and growth factors and environmental factors, and to propose a model suitable for predicting cucumber yield by reflecting these characteristics and to derive important factors.Methods: Using smart farm cucumber data, correlation analysis and dynamic time warping (DTW) are used to analyze the relationship between yield and factors, and three regression models, MLR (multiple linear regression), PLSR (partial least square regression), and SVR (support vector regression), are used for the yield prediction model.Results: The results of this study are as follows; correlation analysis showed that stem thickness and leaf number were highly correlated with yield, and dynamic time warping showed that the increase in the number of nodes and leaf length showed a similar pattern to yield. The relationship between yield and factors can be interpreted differently from the independent influence of a single variable and the perspective of multivariate interaction. In general, environmental management of temperature and humidity during the day plays an important role in improving yield. The SVR model is the most suitable model for predicting cucumber yield because it is advantageous in nonlinear and highly variable data compared to the MLR and PLSR models.Conclusion: This study is expected to expand the applicability of smart farm technology and contribute to optimizing crop growth and improving productivity through data-based predictive modeling.