論文タイトル：Keyword Extraction for Local Foods from Restaurant Menus of Roadside Stations
著者：Yoko Nishihara, Hirofumi Noguchi, and Ryosuke Yamanishi
Some foods are only consumed and available in some regions; these foods are called local foods.
For example, B.C. roll is one of the local foods in Canada.
The local foods are a form of tourism and are expected to attract more tourists.
People in the region are so accustomed to the local foods that they may have difficulty in identifying their values, which means they are unsure whether a particular food is local or not.
If they success to choose appropriate local foods, they may get more tourists.
Therefore, they should know which food should be sold as a local food.
Previous studies have proposed statistical methods for extracting keywords for local foods from restaurant menus on the Internet.
However, the restaurant menus often exclude local foods.
This study applies the previous statistical method to restaurant menus of roadside stations and extracts keywords for their local foods.
These roadside stations are government-designated rest areas located along Japanese roads.
They sell local foods for promotion, while the restaurants provide menus of the local foods.
Thus, if we apply the statistical method to the restaurant menus of roadside stations, we may obtain menus of the local foods.
First, we developed a dataset of restaurant menus of roadside stations.
Then, we apply the previous method to the dataset to extract keywords for the local foods.
We invited participants to our experiment and they evaluated whether the extracted keywords were related to the local foods.
The average rate of keywords for the local foods was 21.1\%.
Furthermore, we discovered that the extracted keywords were not only for foods but also place names, dish names, and their combinations.
書誌情報：Proceedings of the Fifteenth International Conference on Advances in Computer-Human Interactions (ACHI 2022)