TABLE V.

Classification accuracies for adjacent vowel pairs. The displayed scores are the means obtained through 100 repeated runs, along with the 95% CI.

Model	i-ɪ			ɪ-ɛ			ɛ-æ			æ-ɑ
		95% CI			95% CI			95% CI			95% CI
	Mean	Lower	Upper	Mean	Lower	Upper	Mean	Lower	Upper	Mean	Lower	Upper
ALL TOKENS	0.7885	0.7831	0.7953	0.7018	0.6933	0.7123	0.6801	0.6742	0.6867	0.7955	0.7901	0.8025
NO FOCUS	0.7766	0.7708	0.7838	0.6971	0.6894	0.7063	0.6780	0.6708	0.6863	0.7878	0.7829	0.7939
FOCUS	0.8669	0.8651	0.8690	0.7214	0.7144	0.7300	0.6955	0.6914	0.6999	0.8636	0.8618	0.8656