A Unified Model Using Distantly Supervised Data and Cross-Domain Data in NER

Table 4

The performance of using cross-domain data by reinforcement learning method.

Model	Data	EC	News	BC
CD-SHA	Hand + 0% CD	52.65	65.87	61.66
CD-SHA	Hand + 25% CD	62.01	68.98	71.43
CD-SHA	Hand + 50% CD	64.47	71.44	74.07
CD-SHA	Hand + 75% CD	65.94	73.06	78.86
CD-SHA	Hand + 100% CD	65.44	73.51	78.23
CD-RL	Hand +  CD RE	67.46	78.69	78.92