Influencing factors of fruit quality based on robust canonical correlation and partial least squares analysis

Sampling data of soil nutrients, tree characteristics, planting density, fruit nutrition and quality and other indicators from several orchards in one province were used to analyze the influence of each factor on fruit weight．Robust canonical correlation indicated that canonical correlation coefficient increased．Correlation between soil nutrients and fruit nutrients was also analyzed．A multiple regression model with multidimensional response variables between fruit quality and fruit nutrient indicators was established by partial least squares analysis．This study recommends to adjust structure and proportion of soil fertilizer, maintain a reasonable planting density, pay attention to tree pruning and finally focus on fruit quality while pursuing high yield．