Document classification using latent semantic analysis

摘要

Electronically-stored information within a network file system is processed by (a) scanning the information across the network, and identifying each separate file which is suitable for processing; (b) analysing each separate file using latent semantic indexing, comparing the information within a file to information within one or more previously-categorised files, to determine a class to which the file relates; and (c) labelling each separate file with a security grading appropriate to its class, by inserting the security grading into the file, such as a visible watermark, or meta-data associated with the file. Hence information within a network file system is automatically classified, and provided with a security grading.