A Tourist Spot Recommendation Method Based on Image Search and Sentiment Analysis
A Tourist Spot Recommendation Method Based on Image Search and Sentiment Analysis
复制标题
DOI：
10.1109/gcce56475.2022.10014382
复制
发表时间：
2022-10
期刊：
影响因子：
--
通讯作者：
Kouki Oooka;Yuanyuan Wang
中科院分区：
文献类型：
--
作者：
Kouki Oooka;Yuanyuan Wang
提出了一种基于图像搜索的旅游规划和观光图尔斯旅游支持方法，通过计算图像得分和情感得分，基于情感分析推荐满意度高的旅游景点。为了计算图像分数，我们计算用户输入的照片和旅游景点的代表性照片之间的视觉和文本相似性。为了计算情感得分，我们对颜色信息和旅游景点的评论进行情感分析，以推荐满意度高的旅游景点。最后，我们以真实的观光照片资料为例，以影像搜寻为例，对我们所提出的方法进行评估与讨论。
This paper proposes a method for supporting travel planning and sightseeing tours focusing on image search, to recommend tourist spots with high satisfaction based on sentiment analysis by computing image scores and sentiment scores. To compute image scores, we calculate both the visual and textual similarities between user-input photos and representative photos of tourist spots. To compute sentiment scores, we perform a sentiment analysis on the color information and the reviews of tourist spots to recommend tourist spots with high satisfaction. Finally, we evaluated and discussed our proposed method through the image search using the real sightseeing photo data.