多方面属性归一化三维张量模型在区域旅游酒店的推荐应用

为游客个性化推荐理想酒店是旅游质量得以保障的手段之一。首先运用形式化方法将游客、酒店以及游客对酒店的评分从多方面属性综合进行量化与归ー化,得到归一化的酒店因子、游客因子和评分因子;然后,采用一种有偏加权函数计算出一种有偏推荐度,与酒店、游客和评分等因子一起构建了一种以"酒店辟客4平分荐度"四元关系的三维张量模型,有偏推荐度作为张量元素值;最后,采用基于Tucker分解法的算法实现了在高度稀疏的四元关系数据集上按游客分类的有偏性旅游酒店推荐。实验结果表明,采用三维张量模型及算法能实现对高度稀疏的区域旅游酒店数据进行精准旅游酒店推荐,为游客根据个人偏好获得个性化的酒店推荐找出了一种新的方法,有效提高了旅游质量。

A Triple-Dimensional Multi-Aspect Attribute Normalized Tensor Model for Regional Turisthotellet Recommendation

Recommending the ideal hotel for tourists, this is one of the means to ensure the quality of tourism. We Firstly obtain the normalized hotel factor, tourist factor and evaluation factor, by using the formal method to quantify and normalize the tourists, hotels and evaluation with multiple attributes. Then, we define a partial weighting function to calculate the degree of partial recommendation, with the factor of the hotels, tourists and evaluation together to build a triple-dimensional tensor based on 4-tuples relation model including hotel, tourist, evaluation and degree of recommendation, using the degree of recommendation as tensor elements. Finally, we propose a decomposition algorithm based on Tucker to achieve tourism hotel partial recommendation by tourists classified in a higher-sparse 4-tuples relation data set. Experiments show that the triple-dimensional tensor mode and the algorithm can achieve high accuracy for tourist hotel recommendation in higher sparse regional data set, it is based on personal preferences of tourists to get personalized hotel recommended to find a new way to improve the quality of tourism.