Model and Simulation of Maximum Entropy Phrase Reordering of English Text in Language Learning Machine

Table 2

The scale, sort classification, test accuracy, and extracted features of the reranking instance.

Experiment	Total number of instances	Guaranteed order (%)	Reverse order (%)	Others (%)	Test accuracy (%)	Experimental program
1	4353250	81.34	16.48	0.00	92.48	End word features
2	1424597	65.38	32.49	0.00	85.38	Limit the number of instances of order preservation
3	1424597	67.39	32.49	0.00	91.39	First word feature
						Combination features
4	3242782	35.83	20.38	38.49	75.38	Third category
5	3124891	35.38	20.81	38.51	74.30	Extended unaligned words
6	3144248	74.28	34.21	0.00	91.39	Punctuation features