| 来源类型 | Project Reports |
| 规范类型 | 报告 |
| Big data analysis : application to environmental rsearch and service | |
| Sung Won Kang | |
| 发表日期 | 2018-12-31 |
| 出版年 | 2018 |
| 语种 | 英语 ; Korean |
| 摘要 | This report reports the result from second year research of ‘Big Data analysis: Application to Environmental Research and Service’ project. In this project, we try to take advantage of machine learning in Environmental Research. This project consists of three sub-projects. The first one ‘Big Data Environment Research’, experiments machine learning algorithm to environmental research. The second one ‘Big Data Research Infra’ builds up large scale data collection and analysis facility. The third one ‘Big Data Environmental Service’ develops public environmental service using the results from ‘Big Data Environmental Research’ and ‘Big Data Research Infra’. We planned to spend three years for each sub-project, beginning from 2017. 2018 is the second year of first sub-project ‘Big Data Environment Research’. In 2018, we developed four machine learning algorithms ？ CNN algorithm predicting 1~8 hours ahead fine-dust pollution. GRU algorithm predicting 1 week ahead chlorophyl-a pollution. Bidirectional LSTM algorithm for sentiment analysis of climate change SNS data, and Boosted Tree algorithm for analyzing the effect of fine-dust pollution to the number of passengers of Seoul subway. Our sentiment analysis algorithm had 92.95% accuracy. Our CNN algorithm for fine dust pollution prediction cut down RMSE of 1 hour ahead estimation to as low as 2.07？g/m3. Our GRU algorithm for chlorophyl-a pollution prediction had RMSE smaller than the RMSE of Vector Auto Regression by 35.3%. And our Boosted Tree algorithm for subway passenger analysis had RMSE smaller than the RMSE of linear regression by 84.5%. In general, we confirmed that machine learning algorithm had significant advantage in accurate prediction in wide range of environmental research. |
| 来源智库 | Korea Evironment Institute (Republic of Korea) |
| 资源类型 | 智库出版物 |
| 条目标识符 | http://119.78.100.153/handle/2XGU8XDN/198713 |
| 推荐引用方式 GB/T 7714 | Sung Won Kang. Big data analysis : application to environmental rsearch and service. 2018. |
| 条目包含的文件 | ||||||
| 文件名称/大小 | 资源类型 | 版本类型 | 开放类型 | 使用许可 | ||
| ì¬ì _2018_08_ê°ì（9625KB） | 智库出版物 | 限制开放 | CC BY-NC-SA | 浏览 | ||
| 个性服务 |
| 推荐该条目 |
| 保存到收藏夹 |
| 导出为Endnote文件 |
| 谷歌学术 |
| 谷歌学术中相似的文章 |
| [Sung Won Kang]的文章 |
| 百度学术 |
| 百度学术中相似的文章 |
| [Sung Won Kang]的文章 |
| 必应学术 |
| 必应学术中相似的文章 |
| [Sung Won Kang]的文章 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
