联系我们

论文查询结果	返回搜索

论文编号：	14086
作者编号：	2120213273
上传时间：	2023/6/8 22:59:19
中文题目：	基于深度学习模型的电商导购平台内容创作者性格预测及其行为研究
英文题目：	Research on Personality Prediction and Behavior of Content Creators Pertaining to E-commerce Shopping Guide Platform Based on Deep Learning Algorithm
指导老师：	李凯
中文关键字：	MBTI性格特征预测；BERT模型；2SLS；自然语言处理
英文关键字：	BERT Model; natural language processing; 2SLS; MBTI personality prediction
中文摘要：	电商行业规模迅速增加赋予消费者海量消费信息选择权，铺天盖地的商品活动扑面而来，也提高了消费者网购难度和决策时间，商家大额折扣以及消费券等营销活动也难以真正触达有购买意愿的消费者，电商导购平台向消费者提供时效性强的高性价比电商消费科学指南，帮潜在消费者有效缩短了时间成本。其中，内容创作者的力量功不可没，他们通过分享好物和开箱测评等方式提供了主要的内容生产力，引领消费者进行内容消费。电商导购平台通过创作激励机制鼓励内容创作者持续活跃并产出优质消费内容来构建平台坚固的护城河，以期通过双边网络效应吸引更多用户流量，为商家打通营销渠道和销售链路进而创造更多的商业价值。因此，内容导购平台如何持续激发内容创作者的创作积极性和社交积极性是形成其竞争壁垒的关键环节。 本研究旨在从“什么值得买”生活家的评论信息中提取出其性格特征，进而分析预测其行为，结合社会认知理论与社会交换理论，本研究通过预训练半监督学习语言模型BERT对内容创作者性格特征预测、使用工具变量和两阶段最小二乘法对研究假设进行回归检验、使用回归类机器学习算法对内容创作者行为进行扩展性研究预测以及对预训练BERT模型的综合预测能力评估这四个核心步骤对电商导购平台中内容创作者的原创积极性、评论积极性和社交积极性进行了预测分析，研究发现内倾型、直觉型、思维型，以及感知型（INTP）内容创作者比外倾型、感觉型、情感型以及判断型（ESFJ）内容创作者的原创积极性、评论积极性和社交积极性要高。
英文摘要：	The rapid increase in the scale of the e-commerce industry gives consumers the right to choose massive consumption information, and the overwhelming number of commodity activities are coming, which also increases the difficulty and decision-making time of consumers' online shopping. It is also difficult for marketers to really reach consumers who have the desire to purchase. The e-commerce shopping guide platform undertakes the decision-making of their interactive shopping, provides consumers with effective and cost-effective e-commerce consumption scientific guide, and helps potential consumers effectively shorten the time cost. Among them, the power of content creators is indispensable. They provide major content productivity through sharing good things and open-box evaluation, and lead consumers to consume content. The e-commerce shopping guide platform encourages content creators to continue to be active and produce high-quality consumer content through the creation incentive mechanism to build a solid moat of the platform, with a view to attracting more user traffic through the bilateral network effect, opening up marketing channels and sales links for businesses and creating more business value. Therefore, how to continuously stimulate the creative enthusiasm and interactive initiative of content creators is the key link to form its competitive barrier. The purpose of this study is to extract personality characteristics from the "what's worth buying" lifestyle commentaries, and then analyze and predict their behavior. Combining social cognitive theory and social exchange theory, this study uses the semi supervised learning language model BERT to predict content creators' personality characteristics, and uses instrumental variables and two-stage least squares methods to conduct regression tests on research hypotheses, use regression type machine learning algorithms to conduct extensive research and prediction on the behavior of content creators, and to evaluate the comprehensive predictive ability of the pre trained BERT model, the four core steps were used to predict and analyze the creative enthusiasm, the enthusiasm of reviewing, and social initiative of content creators in e-commerce shopping platforms. The study found that introverted, intuitive, thinking, and perceptual (INTP) content creators were more likely to have higher initiative in creative enthusiasm, the enthusiasm of reviewing, and social initiative than extroverted, sensory, feeling and judgmental (ESFJ) content creators
查看全文：	预览  下载(下载需要进行登录)