Training procedure for N-gram-based statistical content classification
View Patent ↗A training procedure for N-gram based statistical document classification has been disclosed. In one embodiment, a set of N-grams is selected out of a second set of N-grams, each of the N-grams having a sequence of N bytes, where N is an integer. Then a statistical content classification model is generated based on occurrences of the N-grams, if any, in a set of training documents and a set of validation documents. The statistical content classification model is provided to content filters to classify content.
1. A computer-implemented method, comprising:
generating a plurality of scores for each of a plurality of N-grams with respect to each of a plurality of categories based on occurrence of the plurality of N-grams in a plurality of training documents;
setting a threshold for each of the plurality of categories based on pre-classification of a plurality of validation documents, occurrence of the plurality of N-grams in the plurality of validation documents, the plurality of scores, and a predetermined false positive limit;
generating a statistical content classification model based on the plurality of scores and the threshold;
assigning a plurality of weights for each of the plurality of N-grams, wherein each of the plurality of weights is associated with a distinct one of the plurality of categories;
storing the statistical content classification model in a data storage device; and
allowing content filters to access the statistical content classification model in the data storage device over a network, wherein the content filters classify contents into one of the plurality of categories using the statistical content classification model.
2. The method of claim 1 , further comprising:
transferring the statistical content classification model to a content filter to be stored thereon, wherein the content filter classifies contents into one of the plurality of categories using the statistical content classification model.
3. A computer-readable storage medium that provides instructions that, if
executed by a processor, will cause the processor to perform operations comprising:
generating a plurality of scores for each of a plurality of N-grams with respect to each of a plurality of categories based on occurrence of the plurality of N-grams in a plurality of training documents;
setting a threshold for each of the plurality of categories based on pre-classification of a plurality of validation documents, occurrence of the plurality of N-grams in the plurality of validation documents, the plurality of scores, and a predetermined false positive limit;
generating a statistical content classification model based on the plurality of scores and the threshold;
assigning a plurality of weights for each of the plurality of N-grams, wherein each of the plurality of weights is associated with a distinct one of the plurality of categories;
storing the statistical content classification model in a data storage device; and
allowing content filters to access the statistical content classification model in the data storage device over a network, wherein the content filters classify contents into one of the plurality of categories using the statistical content classification model.
4. The computer-readable storage medium of claim 3 , wherein the operations further comprise:
transferring the statistical content classification model to a content filter to be stored thereon, wherein the content filter classifies contents into one of the plurality of categories using the statistical content classification model.
5. An apparatus comprising:
a model generator to generate a plurality of scores for each of a plurality of N-grams with respect to each of a plurality of categories based on occurrence of the plurality of N-grams in a plurality of training documents; to set a threshold for each of the plurality of categories based on pre-classification of a plurality of validation documents, occurrence of the plurality of N-grams in the plurality of validation documents, the plurality of scores, and a predetermined false positive limit; to generate a statistical content classification model based on the plurality of scores and the threshold; and to assign a plurality of weights for each of the plurality of N-grams, wherein each of the plurality of weights is associated with a distinct one of the plurality of categories; and
a data storage device coupled to the model generator, to store the statistical content classification model, wherein content filters are allowed to access the statistical content classification model in the data storage device over a network and to classify contents into one of the plurality of categories using the statistical content classification model.
6. The apparatus of claim 5 , wherein the statistical content classification model is transferred to a content filter to be stored thereon, and the content filter classifies contents into one of the plurality of categories using the statistical content classification model.