摘要
在零售行业消费行为分析中,消费者的行为数据往往很难获得。为了获取这些数据,传统的方法是让工作人员跟随顾客并记录数据或者在超市中布置大量的监控摄像头跟踪记录。这些方法虽然可以采集数据用于分析,但是由于需要花费大量的人力物力,而且数据杂乱无章、不易存放、不易修改、不易分析,致使研究结果收效甚微。因此,零售行业急需一种能够快速、准确地采集动线数据并对原始数据进行数据处理、分析的系统。
本文提出了将无线定位技术与零售行业相结合的想法,设计并实现了“基于Wi-Fi无线定位的消费者行为分析系统”。该系统利用Wi-Fi实时定位技术对顾客进行准确、快速的人员定位并记录,代替了以往工作人员人工记录的方法,有效地弥补了记录错误、不易保存、不易分析等缺点,为大型商场提供了便利。经查新,目前尚无无线定位系统与零售行业相结合的应用案例,因此,本文具有应用领域的创新性。
论文作者经过实地调研,研究出了一套处理原始数据的规则。根据这套规则,对定位技术所采集到的原始数据进行优化处理,然后将处理后的数据通过动线重绘,热区显示和报表分析等功能展示给市场营销专家。
论文还提出了新的描述区域吸引力的偏好性模型。新的模型在考虑传统的区域通过次数因素的影响外,还充分考虑了区域停留时间因素和区域购买率因素对区域吸引力的影响。论文给出了此模型下的顾客的偏好动线挖掘算法,该算法可以根据每一步发生区域转移的概率统计得出顾客的偏好动线。
经过实际场景应用的测试,本文实现的系统能够较好地满足零售行业对动线分析的需求,为市场营销专家做出适当的决策支持。
In the previous analysis of consumer behavior, consumers'behavior data was difficult to capture. To capture these data, the traditional methods were to record the data by staff followed the consumer or layout a large number of monitors in the supermarket. These methods could capture the data, but take a great deal of manpower and material resources. And the data was difficult to store and modify. Because the data was disorganized, it was difficult to analyze. So the result was often ineffective. Therefore, the retail industry needed a system which can capture consumers'behavior data quickly and accurately, process and analyze the data.
This paper proposes an idea which is wireless location technology combined with the retail industry and design and implement a Consumers'Behavior Analysis System based on Wi-Fi location technology. It uses Wi-Fi real-time location technology to locate consumers quickly and accurately instead of the previous method of manual recording. It is effective to avoid recording errors. And it becomes easier to store and analyze the data. To the best of our knowledge, there is no case which is wireless location technology combined with the retail industry. Therefore, this paper is innovative.
This paper developed a set of rules for processing the raw data. According to these rules, we optimized the raw data. Then we showed marketing experts the processed data through the shopping paths, the hot zones and the reports.
This paper also proposed a new model of preference to reflect zone attraction. The model not only takes zone visited times into account, but also adds the affection of zone stay time and purchase rate. According to the model, the paper proposes an efficient algorithm which is developed for mining consumers' preferred shopping paths. It can get consumers'preferred shopping paths by the zone-zone transition probabilities statistics in each step. After the test of practical application, the result shows that the system of this paper can meet the demand of the retail industry and give marketing experts decision support.
引文
[1]基于Wi-Fi的实时定位系统(RTLS)[OL]. http://www.rfidinfo.com.cn/Tech/n17 80_1.html.2008.09.18.
[2]ISO发布2.4 GHz RTLS国际标准24730[OL]. http://news.rfidworld.com.cn/20 07_4/20074271152426342.html.2007.04.27.
[3]RTLS的现况与发展趋势(上)[OL]. http://www.rfidinfo.com.cn/Tech/n2092_1.h tml.2009.03.19.
[4]Ekahau发布最新版企业级现场勘测工具[OL]. http://biz.cn.yahoo.com/06022 3/6/g3ur.html.2006.02.23.
[5]Wi-Fi定位系统投入商用号称准确度胜过GPS[OL]. http://tech.sina.com.cn/t/ 2005-06-21/1514641919.shtml.2005.06.21.
[6]AeroScout发布Wi-Fi RFID标签[OL]. http://www.hqew.com/tech/doc/395655 .html.2011.08.31.
[7]调查显示美国实时定位系统市场形势喜人[OL]. http://news.rfidworld.com.cn /2009 2/20092171138558755.html.2009.2.17.
[8]杨娟,郭江涛WiFi通讯技术在煤矿井下的应用[J].煤矿安全.2008(2).
[9]WIFI定位系统在物流园区的多业务应用[OL]. http://www.iotworld.com.cn/ht ml/ImPortLib/201111/b9c6d192c3e0ffcb.shtml.2011.11.29.
[10]顾志峰.基于ZigBee和GPRS的无线网络在煤矿安全中的应用研究[D].华东师范大学:信息学院电子系.2007.
[11]Ergen S C. ZigBee/IEEE 802.15.4 Summary[G].2004.09.10.
[12]雷震洲.支持M2M应用的无线网络技术及发展[J].电信科学.2004.
[13]什么是RFID技术[OL]. http://finance.ifeng.com/stock/special/wlbk/gmxx/2009 0917/1250660.shtml.2009.09.17.
[14]刘鹰,王立辉RFID及RTLS技术和应用[J].应用科技.2004,31(6):51-53.
[15]李要伟,邓成RFID标签激活距离的计算方法[J].实验技术与管理.2006,23(5):31-33.
[16]Zaruba G V, Huber M, Kamangar F A, Chlamtac I. Indoor location tracking using RSSI readings from a single Wi-Fi access point[J]. Wireless networks.2007, 13(2):221-235.
[17]Ekahau, Inc. Ekahau Positioning Engine 4.6 User Guide.2008.
[18]张利.基于WiFi技术的定位系统的设计与实现[D].北京邮电大学:计算机科学与技术学院.2009.
[19]Ekahau, Inc. Deploying RTLS Using Ekahau Site Survey User Guide.2008.
[20]Ekahau, Inc. Ekahau RTLS Guide.2008.
[21]高勇.欧亚卖场动线规划效果调研与建议[D].吉林大学:工商管理学院.2008.
[22]Hui S K, Bradlow E T, Fader P S. Testing behavioral hypotheses using an integrated model of grocery store shopping path and purchase behavior[J]. The Journal of Consumer Research.2009,36(3):478-493.
[23]Hui S K, Fader P S, Bradlow E T. Path data in marketing:An integrative framework and prospectus for model building[J]. Marketing Science.2009,28(2): 320-335.
[24]Larson J S, Bradlow E T, Fader P S. An exploratory look at supermarket shopping paths[J]. International Journal of Research in Marketing.2005,22(4): 395-414.
[25]Hui S K, Bradlow E T, Fader P S. An integrated model of grocery store shopping path and purchase behavior. Working paper, University of Pennsylvania, Philadelphia, http://ssrn.com/abstract=960960.
[26]Hui S K, Fader P S, Bradlow E T. The Traveling Salesman Goes Shopping:The Systematic Deviations of Grocery Paths from TSP Optimality[J]. Marketing Science. 2009,28(3):566-572.
[27]Yada K. String analysis technique for shopping path in a supermarket[J]. Journal of Intelligent Information Systems.2011,36(3):385-402.
[28]Liao I E, Lin W C. RFID Eurasia,2007 1st Annual:Shopping path analysis and transaction mining based on RFID technology, Istanbul, September 5-6,2007[C].
[29]Yaman T, Karabati S, Karaesmen F.20th EURO Mini Conference, Continuous Optimization and Knowledge-Based Technologies:Clustering grocery shopping paths of customers by using optimization-based models, Neringa, May 20-23, 2008[C].
[30]Xing D, Shen J. Efficient data mining for web navigation patterns[J]. Information and Software Technology.2004,46(1):55-63.