Research Article

Gamma-Poisson Distribution Model for Text Categorization

Table 3

Statistical summary of the distribution of nonzero components for the four datasets used here. IQR and SD mean inter quartile range and standard deviation, respectively.

Dataset (%) Median Mean IQR SD

20 Newsgroups 0.05 0.00054 0.00077 0.0005 0.0011
Reuters-21578 1.97 0.00123 0.00167 0.0012 0.0015
Industry Sector 9.04 0.00135 0.00206 0.0019 0.0026
TechTC-100 3.70 0.00276 0.00352 0.0027 0.0037