Association Analysis of Growth Characteristics, WUE, and RUE of Rice in Cold Region Under Different Irrigation Patterns

Abstract

In order to clarify the relationship between growth characteristics of rice, WUE and RUE, the apriori algorithm was used to mine association rules for rice indicators data. In this study, the experiments set up control irrigation, wet irrigation and flood irrigation from May to September in 2017 and 2018, which were carried out in Qing’an Irrigation Experimental Center in the semi-humid area of Heilongjiang Province, China. Because of the difficulty in data collection and the long test period during the study, insufficient data could not be collected for data mining. Therefore, before using the apriori algorithm, the original data was augment by using the generative adversarial network. Correlation analysis of indicators data showed that the correlation among the indicators of rice under the three irrigation patterns was very close. Height, stem diameter and other indicators interacted with each other, which in turn affected the yield, water and radiation use efficiency of rice. Water use efficiency and dry matter got the highest value of correlation analysis under control irrigation was 0.97, yield and dry matter got the second highest value was − 0.95. Yield and height got the highest value of correlation analysis under wet irrigation was 0.94, dry matter and seed setting rate got the second highest value were 0.90, respectively. Yield and height got the highest value of correlation analysis under flood irrigation was − 0.90, stem diameter and dry matter got the second highest value of correlation analysis was 0.85. The results of association analysis combined with three irrigation patterns indicated that the association between seed setting rate, height, leaf area index, dry matter, yield and water use efficiency was the closest. The increase of height, stem diameter, dry matter and seed setting rate can promote rice yield and water use efficiency, while reducing leaf area index and 1000-grain weight will have a negative impact on yield.