摘要
伴随着科技的发展和自2019年以来新型冠状病毒疫情的催化作用,直播逐渐成为一种主流的营销途径之一。然而,对于不同的主播来说,直播间商品的销售情况参差不齐。本研究关注主播的语言风格如何影响直播中不同类别商品的销售。结合现有的理论和分析,本研究搭建了主播的语言风格框架,并运用机器学习模型训练得到一个语言风格分类器。结果显示,相对于享乐品,以商品信息为中心的语言风格对实用品的销售促进作用更大;相反,相对于实用品,以外围信息为中心的语言风格对享乐品的销售促进作用更大。
With the development of technology and the stimuli of COVID-19 pandemic since 2019,e-commerce live streaming is becoming increasingly prevalent as a marketing tool.However,product sales in live streaming vary among different anchors.This study examines the effect of anchors'language styles on product sales in e-commerce live streaming and how this effect differs for utilitarian and hedonic products.Combining the existing theories and our analysis,this study builds a framework for the language styles of anchors and uses the machine learning model to train a language style classifier.The results show that,compared with hedonic products,the product information-centered language styles(authority and guarantee)have a greater promotion effect on sales of utilitarian products;conversely,compared with utilitarian products,the peripheral information-centered language styles(scarcity and liking)has a greater promotion effect on sales of hedonic products.
作者
陈豆豆
白昆龙
井长源
CHEN Doudou;BAI Kunlong;JING Changyuan(Sino-Danish College,University of Chinese Academy of Sciences,Beijing 100190;CAS Research Center on Fictitious Economy&Data Science,Beijing 100190;School of Computer Science and Technology,University of Chinese Academy of Sciences,Beijing 100049)
出处
《科技促进发展》
2023年第3期149-161,共13页
Science & Technology for Development
基金
2021年国家自然科学基金委面上项目(12071458):基于最优化的多视角学习理论、方法与应用研究,田英杰
关键词
电商直播
语言风格
文本分析
支持向量机
机器学习
e-commerce live streaming
language styles
text analysis
SVM
machine learning