| Presentation | 2019-12-04 LSTMの前処理としての特徴選択の有効性 Shiori Koga, Tsunenori Mine, Sachio Hirokawa, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Among RNNs, especially LSTM is capable of long-term memory and can be expected to acquire information including better contextual information, and has been used for various problems including classification problems. In such a classification problem, various attempts have been made to improve the accuracy of classification by selecting features such as words that contribute to classification from among the words that make up sentences. However, since LSTM attaches importance to the context, no attempt has been made to select words in LSTM input as far as we know. In this study, we show the effectiveness of using feature selection as pre-processing of LSTM input, using Word2Vec and BERT as word expressions in LSTM input sentences. The results show that the classification accuracy improves when feature selection is performed. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | text classification / LSTM / feature selection / Word2Vec / BERT |
| Paper # | NLC2019-31 |
| Date of Issue | 2019-11-27 (NLC) |
| Conference Information | |
| Committee | NLC / IPSJ-NL / SP / IPSJ-SLP |
|---|---|
| Conference Date | 2019/12/4(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | NHK Science & Technology Research Labs. |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | The 6th Natural Language Processing Symposium & The 21th Spoken Language Symposium |
| Chair | Takeshi Sakaki(Hottolink) / / Hisashi Kawai(NICT) |
| Vice Chair | Mitsuo Yoshida(Toyohashi Univ. of Tech.) / Kazutaka Shimada(Kyushu Inst. of Tech.) / / Akinobu Ri(Nagoya Inst. of Tech.) |
| Secretary | Mitsuo Yoshida(Ryukoku Univ.) / Kazutaka Shimada(NTT) / / Akinobu Ri(Kyoto Univ.) / (Waseda Univ.) |
| Assistant | Takeshi Kobayakawa(NHK) / Hiroki Sakaji(Univ. of Tokyo) / / Tomoki Koriyama(Univ. of Tokyo) / Yusuke Ijima(NTT) |
| Paper Information | |
| Registration To | Technical Committee on Natural Language Understanding and Models of Communication / Special Interest Group on Natural Language / Technical Committee on Speech / Special Interest Group on Spoken Language Processing |
|---|---|
| Language | JPN-ONLY |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | |
| Sub Title (in English) | |
| Keyword(1) | text classification |
| Keyword(2) | LSTM |
| Keyword(3) | feature selection |
| Keyword(4) | Word2Vec |
| Keyword(5) | BERT |
| 1st Author's Name | Shiori Koga |
| 1st Author's Affiliation | Kyushu University(Kyushu Univ.) |
| 2nd Author's Name | Tsunenori Mine |
| 2nd Author's Affiliation | Kyushu University(kyushu Univ.) |
| 3rd Author's Name | Sachio Hirokawa |
| 3rd Author's Affiliation | Kyushu University(Kyushu Univ.) |
| Date | 2019-12-04 |
| Paper # | NLC2019-31 |
| Volume (vol) | vol.119 |
| Number (no) | NLC-320 |
| Page | pp.pp.13-18(NLC), |
| #Pages | 6 |
| Date of Issue | 2019-11-27 (NLC) |