An Adaptive Point-Of-Interest Recommendation Method Based on Check-in Activity and Temporal-Spatial Probabilistic Models
-
摘要:
针对现有兴趣点(POI)推荐算法对不同签到特征的用户缺乏自适应性问题,该文提出一种基于用户签到活跃度(UCA)特征和时空(TS)概率模型的自适应兴趣点推荐方法UCA-TS。利用概率统计分析方法提取用户签到的活跃度特征,给出一种用户不活跃和活跃的隶属度计算方法。在此基础上,分别采用结合时间因素的1维幂律函数和2维高斯核密度估计来计算不活跃和活跃特征的概率值,同时融入兴趣点流行度来进行推荐。该方法能自适应用户的签到特征,并能更准确体现用户签到的时间和空间偏好。实验结果表明,该方法能够有效提高推荐精度和召回率。
Abstract:Existing Point-Of-Interest (POI) recommendation algorithms lack adaptability for users with different check-in features. To solve this problem, an adaptive POI recommendation method UCA-TS based on User Check-in Activity (UCA) feature and Temporal-Spatial (TS) probabilistic models is proposed. The user check-in activity is extracted using a probabilistic statistical analysis method, and a calculation method of user's inactive and active membership is given. On this basis, one-dimensional power law function and two-dimensional Gaussian kernel density estimation combined with time factor are used to calculate the probability for inactive and active features respectively, and the popularity of POI is incorporated to recommend. This method can adapt to the users' check-in features and reflect the users' check-in temporal-spatial preferences more accurately. The experiments show that the proposed method can effectively improve the recommendation precision and recall.
-
表 1 自适应兴趣点推荐算法(UCA-TS)
输入:签到数据集UCall,推荐的目标用户u,时间槽t; (11) for each l∈L-Lu do 输出:推荐的top-n兴趣点 (12) p1(l|Lu,t’)←1; p2(l|Lu,t’)←0; (1) 使用式(1)计算UCNu,式(3)计算Tu; (13) for each li∈Lu,t’ do (2) k←0; 初始化C={c1, c2}和A={aij}; (14) 计算dli,l和p(l|li); (3) repeat (15) 计算 p1(l|Lu,t’)←p1(l|Lu,t’)×p(l|li); (4) k←k+1; (16) 计算 p2(l|Lu,t’); (5) 更新聚类中心C(k)和A(k-1); (17) end for (6) 更新隶属度矩阵A(k)和聚类中心C(k); (18) 使用式(20)计算pt(l); (7) until 式(5)收敛 (19) end for (8) 返回au={a1u, a2u}; (20) end for (9) for t’=0 to 23 do (21) 使用式(22)计算Pu,t,l; (10) 使用式(21)计算wt’-t; 使用式(15)—式(17)计算H; (22) 排序Pu,t,l并返回top-n兴趣点。 表 2 LBSNs签到数据集的统计情况
数据集 签到数 兴趣点数 用户数 每个用户访问地点平均数 每个地点的平均访问数 每个用户平均签到数 签到密度 Foursquare 194108 5596 2321 46 19 84 0.0149 Gowalla 456905 24236 10162 30 13 45 0.0019 表 3 Foursquare和Gowalla数据集的不活跃和活跃用户数据统计
数据集 用户类别 用户数量 签到记录总数 平均签到记录数 平均签到时间槽数 平均签到地点数 Foursquare 不活跃用户 1908 88181 46 12 28 活跃用户 413 105927 256 19 114 Gowalla 不活跃用户 9756 304621 31 9 21 活跃用户 406 152284 375 18 204 表 4 兴趣点推荐算法在两个数据集上的Fβ指标值(β=1)
数据集 Top-n SK UTE+SE TPR+UM SAMM UCA-TS Foursquare top-5 0.0413 0.0435 0.0555 0.0664 0.0749 top-10 0.0344 0.0356 0.0457 0.0597 0.0669 top-20 0.0245 0.0258 0.0398 0.0520 0.0571 Gowalla top-5 0.0399 0.0425 0.0466 0.0592 0.0663 top-10 0.0313 0.0328 0.0394 0.0551 0.0602 top-20 0.0219 0.0229 0.0283 0.0416 0.0450 -
SI Yali, ZHANG Fuzhi, and LIU Wenyuan. CTF-ARA: An adaptive method for POI recommendation based on check-in and temporal features[J]. Knowledge-Based Systems, 2017, 128: 59–70. doi: 10.1016/j.knosys.2017.04.013 YU Yonghong and CHEN Xingguo. A survey of point-of-interest recommendation in location-based social networks[C]. Trajectory-Based Behavior Analytics: Papers from the 2015 AAAI Workshop, Palo Alto, USA, 2015: 53–60. ZHANG Jiadong and CHOW C Y. CoRe: Exploiting the personalized influence of two-dimensional geographic coordinates for location recommendations[J]. Information Sciences, 2015, 293: 163–181. doi: 10.1016/j.ins.2014.09.014 ZHANG Jiadong and CHOW C Y. Point-of-interest recommendations in location-based social networks[J]. SIGSPATIAL Special, 2015, 7(3): 26–33. doi: 10.1145/2876480.2876486 YUAN Quan, CONG Gao, MA Zongyang, et al. Time-aware point-of-interest recommendation[C]. The 36th International ACM SIGIR Conference on Research and Development in Information Retrieval, Dublin, Ireland, 2013: 363−372. doi: 10.1145/2484028.2484030. ZHANG Jiadong and CHOW C Y. TICRec: A probabilistic framework to utilize temporal influence correlations for time-aware location recommendations[J]. IEEE Transactions on Services Computing, 2016, 9(4): 633–646. doi: 10.1109/TSC.2015.2413783 LI Huayu, GE Yong, HONG Richang, et al. Point-of-Interest recommendations: Learning potential check-ins from Friends[C]. The 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, USA, 2016: 975–984. doi: 10.1145/2939672.2939767. LIU Bin and XIONG Hui. Point-of-interest recommendation in location based social networks with topic and location awareness[C]. The 2013 SIAM International Conference on Data Mining, Austin, USA, 2013: 396–404. doi: 10.1137/1.9781611972832.44. ZHAO Shenglin, ZHAO Tong, YANG Haiqin, et al. STELLAR: Spatial-temporal latent ranking for successive point-of-interest recommendation[C]. The 30th AAAI Conference on Artificial Intelligence, Phoenix, USA, 2016: 315–321. doi: 10.1007/978-981-13-1349-3_5. 任星怡, 宋美娜, 宋俊德. 基于位置社交网络的上下文感知的兴趣点推荐[J]. 计算机学报, 2017, 40(4): 824–841. doi: 10.11897/SP.J.1016.2017.00824REN Xingyi, SONG Meina, and SONG Junde. Context-aware point-of-interest recommendation in location-based social networks[J]. Chinese Journal of Computers, 2017, 40(4): 824–841. doi: 10.11897/SP.J.1016.2017.00824 YIN Hongzhi, CUI Bin, CHEN Ling, et al. Modeling location-based user rating profiles for personalized recommendation[J]. ACM Transactions on Knowledge Discovery from Data, 2015, 9(3): 1–41. doi: 10.1145/2663356 LIU Bin, XIONG Hui, PAPADIMITRIOU S, et al. A general geographical probabilistic factor model for point of interest recommendation[J]. IEEE Transactions on Knowledge and Data Engineering, 2015, 27(5): 1167–1179. doi: 10.1109/TKDE.2014.2362525 LIU Yiding, PHAM T A N, CONG Gao, et al. An experimental evaluation of point-of-interest recommendation in location-based social networks[J]. Proceedings of the VLDB Endowment, 2017, 10(10): 1010–1021. doi: 10.14778/3115404.3115407 YUAN Quan, CONG Gao, and SUN Aixin. Graph-based point-of-interest recommendation with geographical and temporal influences[C]. The 23rd ACM International Conference on Information and Knowledge Management, Shanghai, China, 2014: 659–668. doi: 10.1145/2661829.2661983. 冯浩, 黄坤, 李晶, 等. 基于深度学习的混合兴趣点推荐算法[J]. 电子与信息学报, 2019, 41(4): 880–887. doi: 10.11999/JEIT180458FENG Hao, HUANG Kun, LI Jing, et al. Hybrid point of interest recommendation algorithm based on deep learning[J]. Journal of Electronics &Information Technology, 2019, 41(4): 880–887. doi: 10.11999/JEIT180458 ZHENG Haitao, ZHOU Yingmin, LIANG Nan, et al. Exploiting user mobility for time-aware POI recommendation in social networks[J]. IEEE Access, 2017, 99: 1–14. doi: 10.1109/ACCESS.2017.2764074 LIU Shudong and WANG Lei. A self-adaptive point-of-interest recommendation algorithm based on a multi-order Markov model[J]. Future Generation Computer Systems, 2018, 89: 506–514. doi: 10.1016/j.future.2018.07.008