基于复杂网络的购物篮商品网络分析研究

发布时间：2018-09-15 19:27

【摘要】：超市已经成为人们生活中经常光顾的地方，超市里的商品琳琅满目，商品种类极端丰富，消费者需要处理的信息量急剧增加，已经面临不能从中有效选择所需商品的问题，导致在一个给定的时间内做出不够合理的购买决策。因此，需要超市管理者对消费者进行适当引导，并从中提高超市运营效率。而在超市管理中，商品布货管理和商品定价管理是影响超市成功运营的关键。为了能够科学合理的进行商品布货和定价，需要对商品之间的联系进行分析。也就是说，如果知道哪些商品经常被同时购买，哪些商品的销售会带动其他商品的销售，超市管理者就可以尽可能的把同时购买概率大的商品排放到相近的位置，，或对消费者进行交叉推荐以增加销量，选择那些能产生最大效益的商品进行促销等。因此，超市管理者就需要一个快速有效的分析方法对超市商品属性及相互关系进行分析，得出相应结论，进而可以引导消费者，更重要的是提高超市效益。本文以某大型超市一个阶段内的销售账单为数据来源，利用复杂网络思想，对超市购物篮进行分析研究。主要研究工作有：（1）将超市购物篮问题转化为复杂网络问题，利用复杂网络思想对超市购物篮网络节点度分布、平均路径长度、聚类系数、网络密度以及鲁棒性和脆弱性进行分析研究；（2）对购物篮网络进行商品小团体分析，并从中得到核心商品，对核心商品进行分析研究；（3）利用贪婪算法研究购物篮网络，把购物篮网络划分成多个社团，找出每个社团中的关键商品以及把不同社团联系起来的起桥梁作用的商品，并对这些商品进行分析研究；（4）利用PageRank算法对购物篮商品进行PR值排序，并对所得结果进行合理性解释。
[Abstract]:Supermarkets have become a popular place for people to visit in their daily lives. The supermarket is full of goods, with extremely rich kinds of commodities, and the amount of information that consumers need to deal with increases dramatically. They are already faced with the problem of not being able to select the goods they need effectively from them. Leads to unreasonable purchasing decisions within a given period of time. Therefore, it is necessary for supermarket managers to guide consumers properly and improve the efficiency of supermarket operation. In supermarket management, commodity distribution management and commodity pricing management are the key to the successful operation of supermarket. In order to distribute and price goods scientifically and reasonably, it is necessary to analyze the relationship between commodities. That is to say, if you know which goods are often purchased at the same time, and which goods are sold at the same time, the supermarket manager can, as far as possible, emit the goods that are likely to be purchased at the same time to be close to each other. Or cross-recommend to consumers to increase sales, select the most effective products to promote, and so on. Therefore, supermarket managers need a fast and effective analysis method to analyze the commodity attributes and the relationship between them, and draw the corresponding conclusions, which can guide consumers and, more importantly, improve the efficiency of supermarkets. In this paper, the shopping basket of a large supermarket is analyzed and studied by using the complex network idea, taking the sales bill in one stage of a large supermarket as the data source. The main research works are as follows: (1) the supermarket shopping basket problem is transformed into a complex network problem. The distribution of node degree, average path length and clustering coefficient of supermarket shopping basket network are analyzed by using complex network theory. The network density, robustness and vulnerability are analyzed; (2) the shopping basket network is analyzed by commodity small group analysis, from which core commodities are obtained, and core commodities are analyzed and studied. (3) the greedy algorithm is used to study the shopping basket network. The shopping basket network is divided into several communities to find out the key items in each community and the goods that connect different communities. These items are analyzed and studied. (4) the PageRank algorithm is used to sort the PR value of the shopping basket items, and the results are explained reasonably.
【学位授予单位】：江苏科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F721.7;F274;F224

【参考文献】