Table 7. Classification effectiveness across all ten feature sets using F-measure, ordered by increasing variance.

Feature	Max	Min	Δ	Var
concept + stem	0.9864	0.9085	0.0779	0.000982
concept + stemBigram	0.9825	0.8979	0.0846	0.00104
stem	0.9866	0.901	0.0856	0.00105
concFullBigram	0.895	0.7945	0.1005	0.00135
conceptFull	0.9254	0.8266	0.0988	0.00136
stemBigram	0.9801	0.8767	0.1034	0.00159
concFullMorph + stemBigram	0.9825	0.8766	0.1059	0.00162
concept	0.9526	0.8376	0.115	0.00177
concBigram + stemBigram	0.9171	0.7639	0.1532	0.00313
concBigram	0.9171	0.7639	0.1532	0.00313