【作者】 庞浩；

【导师】 李继云；

【作者基本信息】 东华大学 ， 计算机软件与理论， 2012， 硕士

【摘要】 无线互联网技术的发展,促进了无线定位行业的发展和应用。近年来无线定位技术在医疗、设备管理以及商业智能等领域获得了广泛的应用。在诸多应用中,对大型卖场中顾客动线信息的采集和分析是其重要的一个方面。动线是顾客在超市内购物形成的轨迹,在一定程度上能反映出商场的布局合理性以及顾客的偏好性。动线管理的目标是既满足顾客便捷舒适的购物需求,又能最大程度上促进商家销售任务的完成。动线的通过率、停留率和销售率是动线研究的三项指标,目前对动线研究主要围绕着这些因素展开。动线研究分为数据收集环节和数据分析环节。数据收集分为人工和无线定位技术两种方式,相比前者,无线定位技术具有准确方便和不易侵犯顾客隐私等优点,故而在国内外被广泛应用在有关动线的分析研究领域。本文对动线研究领域中的偏好动线和热区生成算法进行了研究。针对目前基于用户访问矩阵(UAM)和顾客访问矩阵(CAM)的偏好动线求解算法中出现的回路或死循环问题,提出了基于偏好值和簇的偏好图生成(PCG)算法。针对热区研究中存在的热区定义不清、缺乏明确的度量方法和分类算法问题,提出了基于定位信息次数的区域热度的定义,以及基于半径逐级递增的区域定位信息次数统计规则和一种改进型基于k均值的区域热度聚类方法。在以上研究基础上,本文设计并实现了基于WIFI定位技术的超市动线管理系统。用户可选择实时查看顾客的位置信息和移动轨迹,以及动线通过率、停留率和购买率等分析结果。更多还原

【Abstract】 Nowadays the wireless network technology develops rapidly, which promote the development of wireless location service areas. The wireless location technology gains great success in medical measures, instruments managements and business intelligence areas. In business intelligence areas, the research on the shopping path is an important application. The shopping path is defined as the line produced by the customer moving in the shop. The shopping path can show the preferences of the customers towards the different areas. The pass rate, the stay rate and the buy rate is the most three important parameters of the shopping path, and they represent the layout, attraction and sales value of the shop. And these parameters are the center points of the research work. The research of the shopping path can be divided into two paths:data collecting and data analysis. There are two kinds of methods to collect the shopping path data:the manual work and the automatic work. Comparing the former, the latter has nicety, easiness and safe priorities. Considering its low cost and convenience to use, we propose a WIFI deployment model for collecting and analyzing the customers’ shopping path data and purchased items data.In this paper, we do research work of the generation algorithms of preference paths and hot areas. To solve the problems of forever looping paths and break paths which are easily appeared while using the User Access Matrix (UAM) and Customer Access Matrix (CAM) algorithms, we propose a Preference and Cluster based generation Graph (PCG) algorithm. To solve the problems of definition mistiness and absence of a measuring standard and classification of hot areas, we propose a definition of hot degree based on the numbers of location messages, a radius step rising detection algorithm to calculate the location messages, and an improved k-means cluster algorithm for the classification of areas’ hot degrees.On the basis of the above study, we design and implement a Shopping path administration system based on WIFI, which can show the customers’ location information and history tracks, the pass rate, the stay rate,the sales rate and so on.更多还原