期刊
APPLIED INTELLIGENCE
卷 49, 期 3, 页码 858-871出版社
SPRINGER
DOI: 10.1007/s10489-018-1276-1
关键词
Point-of-interest; Location recommendation; Convolutional neural network; Content-aware
资金
- National Natural Science Foundation of China [61772321, 61572301, 61602282, 90612003]
- Natural Science Foundation of Shandong Province [ZR2013FM008, ZR2016FP07]
- Shandong Provincial Key Laboratory of Computer Network [SDKLCN-2016-01]
- Innovation Foundation of Science and Technology Development Center of Ministry of Education
- New H3C Group [2017A15047]
Point-of-interest (POI) recommendation has become an important approach to help people discover attractive locations. But the extreme sparsity of the user-POI matrix creates a severe challenge. To address this challenge, researchers have begun to explore the review content information for POI recommendations. Existing methods are based on bag-of-words or embedding techniques which leads to a shallow understanding of user preference. In order to capture valuable information about user preference, we propose a content-aware POI recommendation based on convolutional neural network (CPC). We utilize a convolutional neural network as the foundation of a unified POI recommendation framework and introduce the three types of content information, including POI properties, user interests and sentiment indications. The experimental results indicate that convolutional neural network is very capable of capturing semantic and sentiment information from review content and demonstrate that the relevant information in reviews can improve POI recommendation performance on location-based social networks.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据