【摘要】 为了解决数字标牌广告投放的推荐问题,研究基于位置数据的推荐算法。在已有的基于矩阵分解思想的兴趣点推荐算法基础上,提出结合矩阵分解思想和商业地理信息数据的兴趣点推荐模型,并在基于位置的数字标牌广告数据上进行实验。实验结果表明,通过为矩阵分解附加商业地理信息的方法,解决了位置访问数据稀疏性的问题,并为数据类型单一,推荐依据不足的问题提供了有效的数据参考及实现方法。为数字标牌广告投放提供了重要的参考依据。更多还原

【Abstract】 To solve the problem of digital signage advertising recommendation,we studied the location data-based recommendation algorithm. On the basis of existing point of interest( POI) recommendation algorithm,which is based on matrix factorisation idea,we proposed a POI recommendation model which combines the matrix factorisation idea and commercial GIS data,and conducted experiments on location-based digital signage advertising data. Experimental results showed that,by the method appending commercial GIS data to matrix factorisation,the problem of location access data sparseness has been solved,and this provides an effective reference and implementation approach for the problems of single data type and insufficient recommendation basis. Our study provides an important reference basis for the digital signage advertising.更多还原