Table 5.

Classification accuracy achieved by two machine learning algorithms on the L > R versus R > L SD group categorization task, using the full lexical feature set (top) and only those features with positive information gain values (bottom).

			Naïve Bayes Gaussian		Naïve Bayes multinomial
		Classification	L > R	R > L	L > R	R > L
All features	Category	L > R (n = 21)	17	4	21	0
		R > L (n = 11)	6	5	4	7
		Accuracy	.68		.88
Selected  features	Category	L > R (n = 21)	20	1	20	1
		R > L (n = 11)	3	8	3	8
		Accuracy	.88		.88