PCA is a tool to reduce the dimensions of factors. It requires me to choose approximate number of eigenvectors