2016 Fiscal Year Annual Research Report
クロスメディア型ＵＧＣマイニングに基づく地域観光知の発見とその利活用に関する研究
| Project/Area Number |
15J01402
|
|---|
| Research Institution | Kyoto University |
|---|
Principal Investigator |
ZHUANG CHENYI 京都大学, 情報学研究科, 特別研究員(DC1)
|
|---|
| Project Period (FY) |
2015-04-24 – 2018-03-31
|
| Keywords | Sightseeing Promotion / Location Discovery / Location Recommendation / Location Revitalization |
|---|
| Outline of Annual Research Achievements |
研究目的: In the sightseeing knowledge mining research, we have three tasks: (1) Location Detection: answering where a sightseeing location is; (2) Popularity Estimation: answering whether a location has been famous or not; and (3) QualityEvaluation: answering whether a location is worth visiting or not. In this year, we have the following achievements.
実績概要: Corresponding to the three tasks, (1) In addition to images, we also collected taxi GPS trajectories in Beijing. New Points of Interests (POIs) are detected using these new data sets; (2) We summarized three proposed models and further devised two new fusing methods. The methods are verified in three cities: Kyoto, Beijing and San Francisco;(3) We improved the environmental psychology based method. We further devised new methods for detecting local culture elements in a city. As a result, a potential sightseeing location's both naturalness and culture are considered for quality evaluation. In a conclusion, our model now can discover potential sightseeing locations in different cities and evaluate their quality from the naturalness and culture aspects.
Our research results are published in two international academic journals, i.e., "Multimedia Tools and Applications" and"The International Journal of Big Data Intelligence".
|
| Current Status of Research Progress |
Current Status of Research Progress
2: Research has progressed on the whole more than it was originally planned.
Reason
According to the plan made in "平成28年度科学研究費助成事業科学研究費補助金交付申請書", we have reached all the goals. Specifically:
(1) We did detect new potential sightseeing locations in Beijing by analyzing new collected taxi GPS trajectory data;
(2) We have made a summary of the three models for location popularity estimation. By devising two new fusing methods, a journal paper is published.
(3) For sightseeing quality evaluation, the current environmental psychology based method is improved and has been published in a journal. Furthermore, by analyzing the images taken in a city, we devised a new method to discover local culture elements in a city. Now, both a location's naturalness and culture are considered when evaluating its quality.
|
| Strategy for Future Research Activity |
In this year, there are mainly three tasks:
(1) Since we have devised a new method for culture elements discovery, for the next step, we should verify its utility in the real world. We are first going to construct a culture element tree for the city: Kyoto. If the results are good, we will try to publish a paper.
(2) After location detection and quality evaluation, the final step is to recommend locations to end-users. To do so, three key relations should be considered, i.e., user-user, user-location, and location-location. Therefore, we are planning to devise a personalized recommendation system, which also considers users' preferences. If everything goes well, we will further focus on how to recommend a sightseeing route rather than a single location.
(3) We are aiming to visualize our research results, by which, all the people (not necessarily being an expert) can understand the research we are doing and the results we obtained.
|
