融合景点季节演变信息的旅游推荐算法  

Tourism recommendation algorithm integrating seasonal evolution information of scenic spots

作　　者：黄士新 左华煜 李慧 於跃成 Huang Shixin;Zuo Huayu;Li Hui;Yu Yuecheng(School of Computer Science,Jiangsu University of Science and Technology,Zhenjiang 212100,China)

机构地区：[1]江苏科技大学计算机学院,江苏镇江212100

出　　处：《南京理工大学学报》2023年第5期651-657,共7页Journal of Nanjing University of Science and Technology

摘　　要：该文在奇异值矩阵分解方法的基础上,提出了一种融合景点季节演变信息的旅游推荐算法。该算法根据景点属性与季节演变之间的关联,将旅游景点的属性划分为静态方面和动态方面,并通过设计包含时间因素的动态偏置函数来刻画用户偏好与景点之间的动态关联。这些静态和动态方面的信息被作为新的偏置项融入有偏奇异分解(Bias singular value decomposition,Bias SVD)模型,以改善用户对旅游景点的评分预测。标准数据集Yelp上的实验结果表明,相比于对用户签到数据无差别对待的推荐方法,该文方法在推荐精度和用户体验方面均有明显的提升。Based on the singular value matrix decomposition method,a tourism recommendation algorithm integrating the seasonal evolution information of tourist spots is proposed in this paper.According to the relationship between attributes of tourist spots and seasonal evolution,attributes of tourist spots are divided into static and dynamic aspects.In this way,with the help of the design of the dynamic bias function including time factors,the dynamic association between user preferences and attractions can be described.Furthermore,these static and dynamic information are incorporated into the Bias SVD model as a new bias item,which improves the user’s score prediction for tourist attractions.The experimental results on the standard data set of Yelp show that compared with the recommendation method that treats user check-in data indiscriminately,recommendation accuracy and user experience has been significiantly improved in our method.

关 键 词：旅游景点推荐 景点热度 评分偏置 景点动态偏置 矩阵分解 协同过滤 社交网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]