| 研究実績の概要 |
研究の目的: We develop a system of mining sightseeing knowledge from public online social communities (e.g., Flickr, Twitter, Facebook, etc.) by the following three tasks. ① Location Detection: answering where a sightseeing location is;② Location Popularity Estimation: answering whether a location has been famous or not; ③ Location Quality Evaluation: answering whether a location is worth visiting or not.
研究実績の概要: In accordance with the research plan I have made in "平成27年度科学研究費助成事業（科学研究費補助金）交付申請書", I have achieved the following results: ① Location Detection: in addition to the locations in Kyoto, I have also collected location information in Beijing when I was serving as an intern in Microsoft Research Asia. Our approach can response to different cities. ② Location Popularity Estimation: for the second task, in accordance with our research plan, we do have implemented a more general Bayesian theory based probabilistic model to do the information asymmetry analysis among different user groups, like locals and foreign tourists. This model is much more robust. ③ Location Quality Evaluation: in addition to the current two quality evaluation methods (i.e., social appreciation andphotographer attention based methods), we devised a whole new method to evaluate a location's scenery quality from the angle of environmental psychology. A novel research sub-topic is proposed.
These research achievements have been published in the international research conferences, i.e., ACM/IEEE ASONAM 2015 and ACM SIGSPATIAL 2015.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
2: おおむね順調に進展している
理由
In this year, I have already reached all the goals in my plan. Specifically:
① We extended our data set about Kyoto to other cities, i.e., Beijing and San Francisco.
② A new robust model is proposed to solve the Location Popularity Estimation problem.
③ A novel sub-topic is proposed in the third Location Quality Evaluation task. About the sub-topic proposed in the third task (i.e., environmental psychology based location quality evaluation), we even have achieved a preliminary research result. It will be presented in the international conference IEEE BigMM 2016, April 20-22, Taipei.
④ I have set up a cooperation relationship with Dr. Xing Xie and Dr. Ruihua Song from Microsoft Research Asia. With their grateful help, some interesting ideas have been exchanged and discussed under my research framework.
|
| 今後の研究の推進方策 |
In this year, there are mainly three tasks:
① Employ human efforts to evaluate our research results. Users' feedback is essential for us to improve the system.
② Improve the new environmental psychology based location quality evaluation methods and apply them to rank more real sightseeing locations (from current 10 locations to 30 or more locations). The number is based on how many subjects we can employ in our experiments.
③ Because a shopping area is an integral part of sightseeing, we will further introduce these areas into our research. To estimate how well a shopping area would attract tourists, some entirely new technologies are expected to be devised this year.
|
