Determining the Dominant Environmental Parameters for Greenhouse Tomato Seedling Growth Modeling Using Canonical Correlation Analysis

摘要

Abstract: This paper tries to determining the dominant environmental parameters for greenhouse tomato seedling growth modeling with canonical correlation analysis (CCA). Although a lot of literatures about crop models have been published, the powerful model for controlling or adjusting the crop growth process in greenhouse is still an open question. In this paper, we use CCA to determining the dominant environmental parameters so that an environmental-parameter-based tomato seedling growth prediction model can be build. To find the dominant environmental parameters, a set of greenhouse tomato environmental-growth parameters will be analyzed with CCA. The canonical correlation coefficients (CCC), the explanation proportion for CCC, and the significance testing values will be calculated firstly. Then CCA will be launched based on the key values. Data analysis results show that CCA can help to interpret the relationship between the environmental parameter set and the growth parameter set. And dominant environmental parameters could be determined based on CCA analysis outputs.