新疆特色农产品线上市场中信任感知、再购买意愿及竞争策略研究

To Promote rapid development of e-commerce is helpful to accelerating the process of turning resource advantage of Xinjiang distinctive farm produce into market advantage, thus boosting the modernization of characteristic agriculture in Xinjiang. Specifically, deep mining of online stock market is significant to pushing the intensive e-commerce development of characteristic products forward in Xinjiang. This project intends to collect trading data from online market of Xinjiang distinctive farm produce and extract quality information from the big sample of micro-data by advanced methods pertaining to data mining, econometrics and statistics. We then measure Internet consumers’ perceived trust in a multi-dimensional way and investigate mechanism of updating and mixing processes of direct and indirect perceived trust, whereby the dynamic evolution of consumers’ trust perception being depicted. Next, dynamic Bayesian network model is constructed to both analyze non-linear causal relations between consumers’ repurchase intention and its antecedents and predict long-term repurchase intention as well. By incorporating Internet consumers’ perceived trust and behavioral intention into game analysis framework, we build a dynamic oligopoly competition model to examine representative vendors’ path of competitive strategy selection and evolutionary game process, the results of which contribute to revealing the dynamic mechanism of online market competition. The research work of this project will not only provide new methods and empirical support for further investigation of online market operating mechanism but also intelligent decision support for upgrading e-commerce application of distinctive farm produce in Xinjiang.

大力发展电子商务可加快新疆特色农产品的资源优势向市场优势转变，促进新疆特色农业的现代化转型。深度挖掘线上存量市场对推进新疆特色农产品电子商务的集约化发展具有重要意义。本项目拟采集新疆特色农产品线上交易的相关信息，运用数据挖掘、计量经济和统计分析方法，从大样本微观数据集中提炼出有效信息。从多个维度衡量消费者的信任感知，研究直接和间接信任感知的更新与融合机制，刻画信任感知的动态演化过程；构建动态贝叶斯网络模型，分析消费者再购买意愿与前置因素的非线性相依关系，预测长期再购买意愿；将消费者信任感知和行为意愿纳入博弈分析框架，建立动态寡头竞争模型，研究典型商家的竞争策略选择和博弈路径，探究线上市场的竞争机制。本项目将为线上市场运行机制的进一步研究提供新思路和实证经验，为新疆特色农产品电子商务发展提供智力决策支持。

本项目以新疆特色农产品线上市场为研究对象，在电子商务经济学理论分析的基础上，综合运用自然语言处理、文本挖掘和深度学习的方法，构建基于在线评论文本进行观点挖掘和意图识别的理论模型，改进并完善在线感知信誉度的量化方法，实证研究感知信誉度对商家经济效益和营销策略的影响作用。从在线评论文本中挖掘出多维度评价信息，定量分析商品质量评价、商家服务评价与再购买意愿之间的动态交互影响作用，进而研究再购买意愿的主要影响因素；在此基础上，设计调查问卷，收集统计调查数据，建立结构方程模型研究在线信任感知与重构意向之间的因果关系；优化动态贝叶斯网络构建算法，并建立了预测长期重购意愿的动态贝叶斯网络模型。从商家策略性行为的视角探究“一价定律”未能实现的原因，实证研究导致价格离散的具体影响因素，挖掘线上市场中商家竞争策略的一般规律性。以定价策略为核心，针对线上真实数据的稀疏性问题和销售期内易逝品价值衰减的突出特点，构建易逝品动态定价模型，优化最优定价算法。在顾客-商品二分网络的基础上，构建基于图注意力网络的电商推荐模型，以提升商品营销的精准性、增强网络商家的市场竞争力。本项目拓展了线上市场运行机制的理论和实证研究，为新疆特色农产品电子商务发展提供智力决策支持。