Emotion Cause Extraction on Chinese Emotion Cause Dataset 10% (test)
70.76PrecisionConvMS-Memnet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ConvMS-Memnetembeddings=pre-trained by skip-grams, number of hops=32017.08 | 70.76 | 68.38 | 69.55 | |
| RBtype=Rule-based2017.08 | 67.47 | 42.87 | 52.43 | |
| Multi-kernel2017.08 | 65.88 | 69.27 | 67.52 | |
| CNNtype=Convolutional Neural Network2017.08 | 62.15 | 59.44 | 60.76 | |
| Memnetembeddings=pre-trained by skip-grams, number of hops=32017.08 | 59.22 | 63.51 | 61.31 | |
| RB+CB+MLtype=Machine Learning, classifier=SVM, features=Rule-based and common-sense features2017.08 | 59.21 | 53.07 | 55.97 | |
| RB+CBtype=Combined rules and common-sense2017.08 | 54.35 | 53.07 | 53.7 | |
| Word2vecclassifier=SVM, features=Word2vec representations2017.08 | 43.01 | 42.33 | 41.36 | |
| SVMfeatures=unigram, bigram, trigram2017.08 | 42 | 43.75 | 42.85 | |
| CBtype=Common-sense based, lexicon=Chinese Emotion Cognition Lexicon2017.08 | 26.72 | 71.3 | 38.87 |