计算机科学
聚类分析
鉴定(生物学)
市场细分
分类
消费者行为
互联网
产品(数学)
数据科学
营销
数据挖掘
万维网
机器学习
人工智能
业务
植物
几何学
数学
生物
出处
期刊:PeerJ
[PeerJ]
日期:2023-08-28
卷期号:9: e1531-e1531
被引量:6
标识
DOI:10.7717/peerj-cs.1531
摘要
The traditional approach to e-commerce marketing encounters challenges in effectively extracting and utilizing user data, as well as analyzing and targeting specific user segments. This manuscript aims to address these limitations by proposing the establishment of a consumer behavior analysis system based on an Internet of Things (IoT) platform. The system harnesses the potential of radio frequency identification devices (RFID) technology for product identification encoding, thus facilitating the monitoring of product sales processes. To categorize consumers, the system incorporates a k-means algorithm within its architectural framework. Furthermore, a similarity metric is employed to evaluate the gathered consumption information and refine the selection strategy for initial clustering centers. The proposed methodology is subjected to rigorous testing, revealing its effectiveness in resolving the issue of insufficient differentiation between customer categories after clustering. Across varying values of k, the average false recognition rate experiences a notable reduction of 20.6%. The system consistently demonstrates rapid throughput and minimal overall latency, boasting an impressive processing time of merely 2 ms, thereby signifying its exceptional concurrent processing capability. Through the implementation of the proposed system, the opportunity for further target market segmentation arises, enabling the establishment of core market positioning and the formulation of distinct and precise marketing strategies tailored to diverse consumer cohorts. This pioneering approach introduces an innovative and efficient methodology that e-commerce enterprises can embrace to amplify their marketing endeavors.
科研通智能强力驱动
Strongly Powered by AbleSci AI