Abstract
Abstract
In the “new economy” represented by the Internet economy background, online consumption as the representative of the new forms of consumption is gradually changing people’s consumption idea and way, online brand loyalty has the extremely important status in the field of online consumption, to stimulate consumption and achieve accurate enterprise marketing, risk control and decision support, improve efficiency and product design business model, business forms, and even change of business thinking, improve enterprise competitiveness is of great significance in the field of online. It is urgent and necessary to apply scientific and effective machine learning method to systematically analyse and study online brand loyalty. In the big data environment, facing the massive data information provided by online consumption, traditional technology methods have gradually failed to meet the competitive needs of enterprises to create and maintain brand loyalty. The traditional random sampling method is difficult to locate the consumer groups with high brand loyalty. At the same time, the traditional data processing technology can not deal with the online consumption behavior with the characteristics of massive, mixed and unstructured data. Traditional approaches have limitations when it comes to the sheer volume of online data and how to use it in real time to target the needs of a brand’s consumer group. The purpose of this study is to build an online consumption of the era of big data model of artificial intelligence, machine learning model, by machine learning method, branded goods purchase behavior of consumers online clustering, building an online brand loyalty measurement model, achieve similar loyalty user clustering, at the same time realize the online measurement of brand loyalty. Among them, focus on machine learning path, machine learning algorithm, model construction methods, and model verification and optimization methods.
Subject
General Physics and Astronomy
Cited by
2 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献