Abstract: A predetermined number of manifest or measurable variables is extracted from each of a plurality of patent documents obtained from a data base or another source and from these manifest variables four or more first order latent variables are defined comprising: knowledge stock, technological scope, international scope and patent value for each patent document and a dependency or causality relationship between the manifest and latent variables and between latent variables is defined. In order to rank the patent documents a PLS Path model operable on the manifest variables is constructed taking into account the latent variables and defined relationships among them, and by applying said PLS Path model the value of the latent variables and the dependency relationship between manifest and latent variables and between latent variables is estimated and the value of the latent variables estimated is used to rank the patent documents.