| Presentation | 2020-12-18 Time-delayed LSTM for historical time series Prediction Rin Adachi, Jun Rokui, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | The research which applies machine learning to the social time series prediction is actively carried out. Among them, many methods are Long Short Term Memory(LSTM), which are a kind of Recurrent Neural Network. We propose a framework to perform social time series prediction by aggregating several social time series data using regression analysis and entering it into a LSTM. LSTM stores the long-term information by the memory cell, but the short-term information is not noticed. In this research, we propose a Recurrent Neural Network architecture for modeling short-term information by delaying historical time-series information. In the verification test, it was confirmed experimentally that the good result was obtained in comparison with LSTM by applying the proposed method. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Time Series Prediction / Long Short Term Memory / LSTM / Historical Time Series / NeuralNetwork |
| Paper # | NC2020-30 |
| Date of Issue | 2020-12-11 (NC) |
| Conference Information | |
| Committee | MBE / NC |
|---|---|
| Conference Date | 2020/12/18(1days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Online |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | |
| Chair | Takashi Watanabe(Tohoku Univ.) / Kazuyuki Samejima(Tamagawa Univ) |
| Vice Chair | Ryuhei Okuno(Setsunan Univ.) / Rieko Osu(Waseda Univ.) |
| Secretary | Ryuhei Okuno(Akita-noken) / Rieko Osu(NTT) |
| Assistant | Akihiro Karashima(Tohoku Inst. of Tech.) / Jun Akazawa(Meiji Univ. of Integrative Medicine) / Ken Takiyama(TUAT) / Nobuhiko Wagatsuma(Toho Univ.) |
| Paper Information | |
| Registration To | Technical Committee on ME and Bio Cybernetics / Technical Committee on Neurocomputing |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Time-delayed LSTM for historical time series Prediction |
| Sub Title (in English) | |
| Keyword(1) | Time Series Prediction |
| Keyword(2) | Long Short Term Memory |
| Keyword(3) | LSTM |
| Keyword(4) | Historical Time Series |
| Keyword(5) | NeuralNetwork |
| 1st Author's Name | Rin Adachi |
| 1st Author's Affiliation | University of Shizuoka(Univ of Shizuoka) |
| 2nd Author's Name | Jun Rokui |
| 2nd Author's Affiliation | University of Shizuoka(Univ of Shizuoka) |
| Date | 2020-12-18 |
| Paper # | NC2020-30 |
| Volume (vol) | vol.120 |
| Number (no) | NC-302 |
| Page | pp.pp.13-18(NC), |
| #Pages | 6 |
| Date of Issue | 2020-12-11 (NC) |