电子商务中基于客户偏好和情感评估的个性化推荐算法
|
摘要
在电子商务时代,在线客户评论已经成为一个普遍的和有价值的顾客和商人做商业决策的信息来源。基于情感评估发现潜在客户的偏好,提出了一种改进的协同过滤算法,并预测顾客对企业服务或产品未来的需求(统称为实体)。具体而言,该方法包括3个主要步骤:层面情感评估、客户偏好挖掘和个性化推荐。首先,情感方面水平评估变换光学字符识别的结构化审查方面水平向量。第二,客户偏好挖掘使用向量从情感中提取层面特征词,并将极性分值分配给每个情感。最后,利用特征词和情感极性评分计算顾客偏好和顾客相似性。根据客户相似性生成服务和产品的个性化推荐。实验结果表明,该方法优于传统的协同过滤方法。
In the era of e-commerce,online customer reviews have become a source of information for the general and valuable customers and businessmen to make business decisions.Based on sentiment evaluation,the paper finds potential customers 'preferences,proposes an improved collaborative filtering algorithm,and predicts customers' future needs for enterprise services or products(collectively referred to as entities).Specifically,the method consists of three main steps:level emotion assessment,customer preference mining and personalized recommendation.First,the level of the structural reviewof the transformation of optical character recognition is evaluated by the level of emotion.Second,customer preference mining uses the vector to extract the feature words from the emotion,and assigns the polarity to each emotion.Finally,the customer preference and customer similarity are calculated by using characteristic words and emotional polarity score.Personalized recommendation of services and products is utterly generated based on customer similarity.The experimental results showthat the method is superior to the traditional collaborative filtering method.
In the era of e-commerce,online customer reviews have become a source of information for the general and valuable customers and businessmen to make business decisions.Based on sentiment evaluation,the paper finds potential customers 'preferences,proposes an improved collaborative filtering algorithm,and predicts customers' future needs for enterprise services or products(collectively referred to as entities).Specifically,the method consists of three main steps:level emotion assessment,customer preference mining and personalized recommendation.First,the level of the structural reviewof the transformation of optical character recognition is evaluated by the level of emotion.Second,customer preference mining uses the vector to extract the feature words from the emotion,and assigns the polarity to each emotion.Finally,the customer preference and customer similarity are calculated by using characteristic words and emotional polarity score.Personalized recommendation of services and products is utterly generated based on customer similarity.The experimental results showthat the method is superior to the traditional collaborative filtering method.
引文
[1]张昊,纪宏超,张红宇.XGBoost算法在电子商务商品推荐中的应用[J].物联网技术,2017,7(2):102-104.
[2]王许辉,吕肖飞.电子商务个性化推荐系统模型介绍[J].信息与电脑(理论版),2017(10):85-89.
[3]杨志平,徐骞,马铭.基于中间商品兴趣度加权的内容推荐算法[J].北华大学学报(自然科学版),2017,18(6):831-834.
[4]叶建龙.基于Mahout的个性化电子商务推荐系统研究[J].信息通信,2017(10):129-130.
[5]陈星星.基于多最小支持度关联规则的电子商务推荐系统的研究与实现[D].广州:暨南大学,2017.
[6]张光.基于离群数据挖掘的电子商务推荐系统研究[J].自动化与仪器仪表,2017(8):21-22,25.
[7]吴金李,张建明.基于二分K-means的协同过滤推荐算法[J].软件导刊,2017,16(1):26-29.
[8]余明艳,郁春兰.农产品电子商务推荐系统设计与实现[J].经济研究导刊,2017(28):21-22.
[9]战凯.基于遗忘函数和项目流行度的推荐算法研究[D].银川:北方民族大学,2017.
[10]张航,叶东毅.一种基于多正则化参数的矩阵分解推荐算法[J].计算机工程与应用,2017,53(3):74-79.
[11]王茜,王均波.一种改进的协同过滤推荐算法[J].计算机科学,2010,37(6):226-228,243.
[12]文俊浩,袁培雷,曾骏,等.基于标签主题的协同过滤推荐算法研究[J].计算机工程,2017,43(1):247-252,258.
[2]王许辉,吕肖飞.电子商务个性化推荐系统模型介绍[J].信息与电脑(理论版),2017(10):85-89.
[3]杨志平,徐骞,马铭.基于中间商品兴趣度加权的内容推荐算法[J].北华大学学报(自然科学版),2017,18(6):831-834.
[4]叶建龙.基于Mahout的个性化电子商务推荐系统研究[J].信息通信,2017(10):129-130.
[5]陈星星.基于多最小支持度关联规则的电子商务推荐系统的研究与实现[D].广州:暨南大学,2017.
[6]张光.基于离群数据挖掘的电子商务推荐系统研究[J].自动化与仪器仪表,2017(8):21-22,25.
[7]吴金李,张建明.基于二分K-means的协同过滤推荐算法[J].软件导刊,2017,16(1):26-29.
[8]余明艳,郁春兰.农产品电子商务推荐系统设计与实现[J].经济研究导刊,2017(28):21-22.
[9]战凯.基于遗忘函数和项目流行度的推荐算法研究[D].银川:北方民族大学,2017.
[10]张航,叶东毅.一种基于多正则化参数的矩阵分解推荐算法[J].计算机工程与应用,2017,53(3):74-79.
[11]王茜,王均波.一种改进的协同过滤推荐算法[J].计算机科学,2010,37(6):226-228,243.
[12]文俊浩,袁培雷,曾骏,等.基于标签主题的协同过滤推荐算法研究[J].计算机工程,2017,43(1):247-252,258.