Training procedure for N-gram-based statistical content classification

摘要

Training step is used to be disclosed based on the classification of n-gram statistical doucument, export. In one embodiment, second of n-gram is selected in one group of n-gram, each n-gram has N number of byte of a sequence, and wherein N is integer. Then statistical content is generated disaggregated model and is talked about based on the n-gram of appearance, one group of Training document and one group of verifying document. The statistical content disaggregated model is provided to content filter and carrys out categorised content.