MIT Press

Table 4:

Performance comparison of models using full texts and our implemented models using paper abstracts (and authors and research areas as categories for basis-customized models) as inputs in the AAPR data set.

Models	Accuracy
using full text (Yang et al., 2018)
LSTM	60.5
MHCNN	67.7
using abstract and categories (our setting)
LSTM	60.6
MHCNN	63.7
BiLSTM	61.7
BiLSTM +word-basis-cust	65.8
BiLSTM +attention-basis-cust	65.9
BiLSTM +encoder-basis-cust	66.2

Models	Accuracy
using full text (Yang et al., 2018)
LSTM	60.5
MHCNN	67.7
using abstract and categories (our setting)
LSTM	60.6
MHCNN	63.7
BiLSTM	61.7
BiLSTM +word-basis-cust	65.8
BiLSTM +attention-basis-cust	65.9
BiLSTM +encoder-basis-cust	66.2

View Large

This Feature Is Available To Subscribers Only