基于标签的个性化旅游推荐

发布时间：2019-01-07 07:31

【摘要】：旅游景点数量庞大,而用户本身旅游次数有限,所以用户旅游数据非常稀疏,进而影响了推荐结果的准确度.为了解决这一问题,从海量游记中提取与旅游景点密切相关的4个因素——地域、时间、主题、类型相关的特征标签,来丰富数据信息.一方面通过基于标签内容的方法为用户推荐感兴趣的景点;另一方面,用景点特征标签描述用户兴趣特征,根据用户兴趣标签找到相似用户群,通过协同过滤的方法为用户推荐感兴趣的景点.实验结果表明,基于标签的协同过滤算法较传统的协同过滤算法推荐准确率提高了63.7%,比基于景点热度的推荐算法提高了22.5%;基于标签内容的推荐算法比基于景点热度的推荐算准确率提高了27.6%.进一步,通过线性加权的方式混合两种算法,使两种算法优势互补,从而得到更好的推荐效果.最终使得基于标签的混合算法的准确率,比基于标签的协同过滤算法提高了61.3%,比基于标签内容的推荐算法提高了54.7%.旅游景点推荐准确度的提高,将带来更好的用户体验,使在线旅游网站更加具有竞争力.
[Abstract]:Because of the large number of tourist attractions and the limited number of trips by the users themselves, the travel data of the users are very sparse, thus affecting the accuracy of the recommended results. In order to solve this problem, four factors closely related to tourist attractions, such as region, time, theme and type, are extracted from mass travel notes to enrich data information. On the one hand, through the method based on label content to recommend the interesting scenic spots for the user; On the other hand, the feature label is used to describe the user's interest, and the similar user group is found according to the user's interest label, and the interest of the user is recommended by collaborative filtering method. The experimental results show that the recommendation accuracy of the tag based collaborative filtering algorithm is 63.7 higher than that of the traditional collaborative filtering algorithm, and 22.5% higher than that of the recommendation algorithm based on the heat of scenic spots. The accuracy of recommendation algorithm based on label content is 27.655% higher than that based on heat of scenic spot. Furthermore, by combining the two algorithms with linear weighting, the advantages of the two algorithms are complementary and the recommended results are better. Finally, the accuracy of the tag-based hybrid algorithm is improved 61.3% compared with the tag-based collaborative filtering algorithm, and 54.7% higher than the tag-based content recommendation algorithm. The improvement of recommendation accuracy will bring better user experience and make online travel website more competitive.
【作者单位】：中山大学数据科学与计算机学院;
【基金】：国家自然科学基金(61502543) 广东省自然科学基金杰出青年项目(2016A030306014);广东省自然科学基金博士启动项目(2014A030310180) 中央高校基本科研业务费专项(16lgzd15)资助
【分类号】：TP391.3

【相似文献】