Provide results for CNN+LSTM+CRF for cQA answer tagging by the paper:
http://www.aclweb.org/anthology/C/C16/C16-1117.pdf
It is published in COLING 2016.
The experiments are carried on SemEval 2015 task-3 cQA dataset.
----ARC-I: 58.96 on F1
----ARC-II: 58.29 on F1
The systems can be easily implemented with the help of https://github.com/glample/tagger and the parameter settings described in the coling paper.
Any work compared with this paper, please cite
@inproceedings{Xiang2016Incorporating, title={Incorporating Label Dependency for Answer Quality Tagging in Community Question Answering via CNN-LSTM-CRF}, author={Xiang, Yang and Zhou, Xiaoqiang and Chen, Qingcai and Zheng, Zhihui and Wang, Xiaolong and Qin, Yang}, booktitle={COLING}, year={2016}, }