Hi,
I'm not sure but I think there is an error in the theory of PCA.
At the step of the computation of influence score, the graph used is the graph with the gene C modified. However, the dataset showed is still with the original values, so C is not corresponding, neither the lines of PC, thus the influence scores does not fit with the original data. There is also the wrong graph for the rotation.
Thank you !
Hi, I'm not sure but I think there is an error in the theory of PCA. At the step of the computation of influence score, the graph used is the graph with the gene C modified. However, the dataset showed is still with the original values, so C is not corresponding, neither the lines of PC, thus the influence scores does not fit with the original data. There is also the wrong graph for the rotation. Thank you !
Hortense