[Canonical Correlation and Redundancy Analysis on the Indices of Abdominal Ultrasonography of Schistosomiasis Japonica].

Yi-biao Zhou,Gen-ming Zhao,Qing-wu Jiang,An Nin,Jian-guo Wei,Zu-lu Gao,Qi-long Wu
DOI: https://doi.org/10.3969/j.issn.1000-7423.2004.01.007
2004-01-01
Abstract:OBJECTIVE:To explore the correlation of the ultrasound indices of liver and spleen in schistosomiasis japonica and with infection frequency, infection time and EPG. METHODS:The canonical correlation was applied to analyze the correlation of the hepatosplenic ultrasound indices in schistosomiasis japonica, and the correlation of the ultrasound indices with infection frequency, infection time and EPG. The proportions of variation related to each other in liver and spleen were analyzed by the redundancy analysis. RESULTS:The correlation coefficients of the first pair correlation canonical variable were 0.7842, 0.5483 and 0.5800, 0.4220, respectively, in males and females without infection, males and females with infection history (P<0.01). The correlation coefficients of the first pair correlation canonical variable were 0.6063, 0.5215 and 0.6595, 0.3849, respectively, in male negatives and female negatives, male positives and female positives (P<0.01). In groups of males and females without infection, the variations of liver ultrasound indices related with the variation of spleen ultrasound indices were 43.5% and 17.5% respectively, and in groups infection history, they were 22.1% and 11.4% respectively. In male and female negative groups, the variations of liver ultrasound indices related with the variation of spleen ultrasound indices were 26.8% and 16.8% respectively, and in positive groups, they were 27.6% and 10.7% respectively. The infection frequency, infection time and EPG in the stool-positive groups were not significantly related with the canonical variables of liver and spleen ultrasound indices (P>0.05). CONCLUSION:There is a significant canonical correlation between the liver and spleen ultrasound indices in schistosomiasis japonica, especially in males. In groups with infection history, the infection frequency, infection time and EPG of those stool-positives are not relevant to the canonical variables of liver and spleen ultrasound indices.
What problem does this paper attempt to address?