A Sentiment-aware Delightful Walking Route Recommendation System Considering the Scenery and Season

• Published in: IEEE ... International Conference on Data Mining workshops pp. 867 - 872
• Main Authors: Li, Da, Ishitsubo, Shiho, Yamauchi, Katsuyuki, Siriaraya, Panote, Nakajima, Shinsuke, Kawai, Yukiko
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2021
• Subjects: Feature extraction; Internet; Legged locomotion; Mood; route recommendation system; scenery and season; Sentiment analysis; social media; walking support
• ISSN: 2375-9259
• DOI: 10.1109/ICDMW53433.2021.00112

With the aggravation of the aging society, people choose exercise by walking increasingly for maintaining health or losing weight. However, even if we understand that exercising regularly has many benefits for our body and brain, it is not easy to keep on it. Therefore, the research of walking support system is of great significance. In order to make the walking route more interesting, in this paper, we propose a sentiment-aware delightful walking route recommendation system considering the scenery and season. Firstly, we analyze the content of tweets, calculate the emotional score by utilizing sentiment analysis tools, and extract the related words of season and sentiment features. Secondly, we calculate the cosine similarity of related words and geotagged tweets, and recommend the walking routes using the location of the tweets. Finally, to verify the validity of our walking route recommendation system, we ask 41 test subjects to evaluate our system by watching videos of walking routes, which are generated by tweet contents and Google Street View.