应用于电子商务环境的商业模式挖掘和预测方法（英文）
发布时间：2019-04-12 20:24
【摘要】：目的:随着电子商务的发展,商品交易形式发生了翻天覆地的变化。在这种新型虚拟交易平台上,消费者快速而又广泛地浏览、购买、评价各种价廉物美的商品,而商品也同样地进行着产品创新、营销拓展、物流改善。所以,消费者和商品之间早已不是简单的一对一直接买卖关系,而是消费者和消费者之间的社交网络,消费者和商品之间的交易网络构成电子商务中最重要的物质载体。于是,对于商品销量预测这一传统经济问题,在电子商务的大环境下有了新型的研究意义。为了能有效和深入地研究在电子商务环境下的商品销量发展模式,本文的目的在于提供一种应用于电子商务的商品销量预测算法,该算法注重于从消费者社会影响力分析入手,更好地适应真实电子商务环境中的销量预测需求。创新点:首先,本文算法不仅考虑到消费者自身的特征,同时还考虑到存在于消费者之间的社会影响力,考虑到在真实的电子商务中,消费者之间传递商品的价格信息或评价信息十分便捷,因而本文算法很好地切合了实际的应用环境。其次,本文算法定义了交易环境中的两种社会影响力,即"同一商品中消费者互相作用产生的影响力"和"不同商品之间消费者互相作用产生的影响力",分别考虑到单一商品的交易环境和多个商品互相作用的交易环境中消费者行为,其中以上两种社会影响力都是由真实消费者社交网络分析提炼得来的,使得本算法更加切合真实交易网络的内在结构。方法:本文算法将商品销量分为主体部分和噪声部分,很好地模拟了真实交易环境中,商品销量的构成是受多成分影响的。并且在预测模型中,对主体部分和噪声部分分别设置了不同的约束条件,具体为要求商品销量的主体部分在时间上应该保持平滑性,并要求商品销量的噪声部分是稀疏的,以上两个约束很好地反映了真实交易环境中商品销量的变化形式。结论:本文研究电子商务环境下商品销量的发展模式,并提出描述消费者之间关系的两种社会影响力网络。将此社会影响力网络整合入商品销量构成模型中,最后提出对这些商品销量的预测算法。特别是通过在真实的数据环境中(阿里巴巴女装数据)进行算法测试,并结合与传统销量预测算法的比较,展示在复杂数据环境下本算法的有效性。
[Abstract]:Objective: with the development of e-commerce, the form of commodity trade has changed dramatically. On this new virtual trading platform, consumers browse, purchase and evaluate all kinds of cheap and beautiful goods quickly and extensively, and the goods also carry on product innovation, marketing expansion and logistics improvement. Therefore, the relationship between consumers and goods is not a simple one-to-one direct trading relationship, but a social network between consumers and consumers, and the transaction network between consumers and goods constitutes the most important material carrier in e-commerce. Therefore, for the traditional economic problem of commodity sales forecast, it has a new research significance under the environment of e-commerce. In order to effectively and deeply study the development mode of commodity sales under the environment of e-commerce, the purpose of this paper is to provide an algorithm for forecasting the sales volume of goods applied to e-commerce. The algorithm focuses on the analysis of consumers' social influence. Better adapt to the real e-commerce environment sales forecast demand. Innovation: first of all, this algorithm not only takes into account the characteristics of consumers, but also considers the social influence between consumers, taking into account that in real e-commerce, It is very convenient for consumers to transfer the price information or evaluation information of goods, so the algorithm in this paper fits well with the practical application environment. Secondly, the algorithm defines two kinds of social influence in the trading environment, that is, "the influence of consumer interaction in the same commodity" and "the influence of consumer interaction between different commodities". Considering the trading environment of a single commodity and the behavior of consumers in a trading environment in which multiple goods interact, the above two kinds of social influence are extracted from the analysis of real consumer social networks. It makes this algorithm more suitable to the internal structure of the real trading network. Methods: this algorithm divides the product sales into the main part and the noise part, and well simulates the real trading environment, the composition of the commodity sales is affected by the multi-component. And in the prediction model, different constraints are set on the main part and the noise part, specifically, the main part of the commodity sales should be smooth in time, and the noise part of the product sales is sparse, and the main part of the product sales should be smooth in time, and the noise part of the product sales should be sparse. The above two constraints well reflect the changing form of commodity sales in the real trading environment. Conclusion: this paper studies the development model of commodity sales in e-commerce environment, and puts forward two kinds of social influence networks which describe the relationship between consumers. This network of social influence is integrated into the model of commodity sales, and finally, the prediction algorithm of these commodity sales is put forward. Especially, the algorithm is tested in the real data environment (Alibaba women's dress data) and compared with the traditional sales forecasting algorithm to show the effectiveness of the algorithm in the complex data environment.
【作者单位】： Department
【基金】：supported by the National Basic Research Program(973)of China(No.2012CB316400) Zhejiang University-Alibaba Financial Joint Lab,Zhejiang Provincial Engineering Center on Media Data Cloud Processing and Analysis,China the US National Science Foundation(No.CCF-1017828)
【分类号】：F724.6;TP301.6
[Abstract]:Objective: with the development of e-commerce, the form of commodity trade has changed dramatically. On this new virtual trading platform, consumers browse, purchase and evaluate all kinds of cheap and beautiful goods quickly and extensively, and the goods also carry on product innovation, marketing expansion and logistics improvement. Therefore, the relationship between consumers and goods is not a simple one-to-one direct trading relationship, but a social network between consumers and consumers, and the transaction network between consumers and goods constitutes the most important material carrier in e-commerce. Therefore, for the traditional economic problem of commodity sales forecast, it has a new research significance under the environment of e-commerce. In order to effectively and deeply study the development mode of commodity sales under the environment of e-commerce, the purpose of this paper is to provide an algorithm for forecasting the sales volume of goods applied to e-commerce. The algorithm focuses on the analysis of consumers' social influence. Better adapt to the real e-commerce environment sales forecast demand. Innovation: first of all, this algorithm not only takes into account the characteristics of consumers, but also considers the social influence between consumers, taking into account that in real e-commerce, It is very convenient for consumers to transfer the price information or evaluation information of goods, so the algorithm in this paper fits well with the practical application environment. Secondly, the algorithm defines two kinds of social influence in the trading environment, that is, "the influence of consumer interaction in the same commodity" and "the influence of consumer interaction between different commodities". Considering the trading environment of a single commodity and the behavior of consumers in a trading environment in which multiple goods interact, the above two kinds of social influence are extracted from the analysis of real consumer social networks. It makes this algorithm more suitable to the internal structure of the real trading network. Methods: this algorithm divides the product sales into the main part and the noise part, and well simulates the real trading environment, the composition of the commodity sales is affected by the multi-component. And in the prediction model, different constraints are set on the main part and the noise part, specifically, the main part of the commodity sales should be smooth in time, and the noise part of the product sales is sparse, and the main part of the product sales should be smooth in time, and the noise part of the product sales should be sparse. The above two constraints well reflect the changing form of commodity sales in the real trading environment. Conclusion: this paper studies the development model of commodity sales in e-commerce environment, and puts forward two kinds of social influence networks which describe the relationship between consumers. This network of social influence is integrated into the model of commodity sales, and finally, the prediction algorithm of these commodity sales is put forward. Especially, the algorithm is tested in the real data environment (Alibaba women's dress data) and compared with the traditional sales forecasting algorithm to show the effectiveness of the algorithm in the complex data environment.
【作者单位】： Department
【基金】：supported by the National Basic Research Program(973)of China(No.2012CB316400) Zhejiang University-Alibaba Financial Joint Lab,Zhejiang Provincial Engineering Center on Media Data Cloud Processing and Analysis,China the US National Science Foundation(No.CCF-1017828)
【分类号】：F724.6;TP301.6
【共引文献】
相关期刊论文 前10条
1 曾伏娥;代婷婷;朱妮亚;;网络社区成员回应社区广告的社会性影响因素研究[J];管理学报;2013年08期
2 王永贵;马双;;虚拟品牌社区顾客互动的驱动因素及对顾客满意影响的实证研究[J];管理学报;2013年09期
3 范钧;付沙沙;汤锦旦;;虚拟品牌社区互动对顾客公民行为的影响[J];商业研究;2014年02期
4 李雪欣;包文莉;夏天;;虚拟品牌社区顾客动机对产品创新行为的影响[J];东北大学学报(社会科学版);2014年05期
5 胡磊;高迎;;微博社区成员参与动机实证研究[J];北京大学学报(自然科学版);2014年05期
6 胡磊;王琼;范p，
本文编号：2457365
本文链接：https://www.wllwen.com/guanlilunwen/yingxiaoguanlilunwen/2457365.html
