统计与数据科学学院知识库

  农村电子商务作为“三农”发展的重要抓手，既可以促进农产品上行，提升农产品流通效率，又可以帮助农村地区优化产业结构，推动农业升级、农村发展、农民增收，是我国全面推进乡村振兴战略的重要实践选择。农村电子商务持续助力“三农”发展，主要表现为电商助农模式的运行。在电商助农模式下，通过电子商务平台，将农产品与消费者直接连接起来，有效解决了传统农产品销售的中间环节问题，同时为农产品提供了更多元化的销售渠道，使得农产品能够更广泛地进入市场，得到消费者的认可和购买。但是目前电商助农的发展过程中出现了一些关键问题，如服务质量低、消费者需求未得到满足等导致农产品缺乏产品竞争力，使得其销售滞后。而产品竞争力是企业或商家发展的核心，率先满足消费者的需求是产品竞争力最直接的价值体现，与电商助农产业的可持续发展息息相关。因此，需要从消费者需求出发，对电商助农产品运营策略进行改进。

  本文以京东平台电商助农水果产品为研究对象，设计了电商助农水果产品运营策略改进方案，为电商助农水果产品商家提供产品运营策略改进决策支持。首先，通过八爪鱼采集器在京东平台爬取了多家助农水果产品商家和一般水果产品商家消费者在线评论数据，对其进行评论去重、文本去噪等文本数据预处理操作；其次，构建水果产品领域属性词典，确定产品评价维度即产品属性，基于PaddleNLP搭建SKEP模型进行属性级情感分析，从在线评论数据中提取出相应的属性词、观点词及情感倾向，再利用基于情感词典的方法计算水果产品各项属性的情感得分；之后，利用属性级情感得分结果计算助农水果产品商家和一般水果产品商家水果产品的属性绩效表现、属性重要性值，得出电商助农水果产品整体的市场表现；确定电商助农水果产品目标研究商家，通过竞争者识别模型识别其竞争对手，再通过IPCA和DIPCA模型分析得出目标商家水果产品的优劣属性；最后，针对目标商家第一优先级改进劣势属性，通过文本挖掘技术TF-IDF对此属性的原始负面评论提取消费者不满意因素的热词，利用K-Means聚类算法对消费者不满意因素的热词进行聚类分析，最终以热图形式将运营策略改进方案形成依据进行可视化展示，提出具体的改进方案。

  研究得出目标商家的“果面”属性为第一优先级改进劣势属性，应该采取的产品改进方案为：一方面，重点改进“果面”属性，即严控水果采摘保鲜环节、严控筛检环节、改善包装以及选择可靠的物流公司；另一方面，还需要根据属性改进收益合理分配资源投入其他属性改进。通过对全文工作和结果的总结，得出了以下研究结论：第一，电商助农水果产品的消费者满意度较低且市场表现相对较差，在电商水果市场上的竞争力较弱；第二，消费者的偏好和需求可能会随着时间的推移而变化，电商助农产品商家应及时关注消费者的偏好和需求，并适时调整运营策略和资源配置；第三，消费者满意度越高的产品属性其竞争力越强，且竞争对手对产品属性的改进可能会削弱自身产品的竞争力，因此电商助农产品商家还应不断完善自己的产品，避免陷入“不进则退”的陷阱。

  根据研究结论，对电商助农不同主体提出了不同的建议：对消费者来说，一是积极对商品进行真实的评价，二是京东平台是水果等生鲜产品和支持消费助农的不错选择，三是购买水果等生鲜产品要了解清楚退换赔付情况。对电商助农水果商家来说，一是严格控制产品质量，二是完善自身服务，三是合理分配资源投入，四是紧跟消费需求、把握消费趋势。对于平台来说，一是优化物流体系建设，二是完善包装管理，三是健全售后服务体系。

      As an important grasp of the development of agriculture, rural areas and farmers, rural e-commerce can not only promote the upward of agricultural products and improve the circulation efficiency of agricultural products, but also help rural areas optimize the industrial structure, promote agricultural upgrading, rural development and farmers' income increase. It is an important practical choice for China to comprehensively promote the rural revitalization strategy. Rural e-commerce continues to help the development of agriculture, rural areas and farmers, mainly reflected in the operation of the e-commerce helping agricultural mode. Under the e-commerce helping agricultural mode, through the e-commerce platform, agricultural products are directly connected with consumers, which effectively solves the problem of the intermediate link of traditional agricultural product sales, and provides more diversified sales channels for agricultural products, so that agricultural products can enter the market more widely and be recognized and purchased by consumers. However, there are some key problems in the development process of e-commerce helping agriculture, such as low service quality and unmet consumer demand, which lead to the lack of product competitiveness of agricultural products, making their sales lag behind. The product competitiveness is the core of the development of enterprises or businesses. To meet the needs of consumers first is the most direct embodiment of the value of product competitiveness, which is closely related to the sustainable development of e-commerce helping agriculture industry. Therefore, it is necessary to improve the operation strategy of e-commerce helping agricultural products from the perspective of consumer demand.

      Taking the e-commerce fruit products on JD platform as the research object,this paper designs the improvement plan of e-commerce fruit product operation strategy,and provides decision support for the improvement of product operation strategy for e-commerce fruit product merchants.Firstly,the online review data of several e-commerce fruit product merchants and general fruit product merchants are crawled on Jingdong platform through octopus collector,and the text data preprocessing operations such as deduplication and text denoising are performed on the review.Secondly,the attribute dictionary of fruit product field is constructed to determine the product evaluation dimension,namely product attributes.The SKEP model is built based on PaddleNLP for Aspect(Attribute)-based Sentiment Analysis,and the corresponding attribute words,view words and sentiment orientation are extracted from the online review data.Then the attribute-based sentiment score of fruit product attributes is calculated using the attribute-based sentiment score method.Then,the attribute performance and attribute importance value of fruit products of e-commerce fruit product merchants and general fruit product merchants are calculated using the attribute-level sentiment score results,and the overall market performance of e-commerce fruit products is obtained.The target merchants of e-commerce fruit products are determined,and their competitors are identified through the competitor identification model.Then the advantages and disadvantages of fruit products of target merchants are analyzed through IPCA and DIPCA models. Finally, aiming at the first priority of the target merchants to improve the disadvantage attributes, the hot words of consumer dissatisfaction factors are extracted from the original negative comments of this attribute by the text mining technology TF-IDF, and the hot words of consumer dissatisfaction factors are clustered by the K-Means clustering algorithm. Finally, the basis of the formation of the operation strategy improvement plan is visualized in the form of heat map, and the specific improvement plan is proposed.

      The study concludes that the "fruit surface" attribute of the target merchants is the first priority of the disadvantage attributes, and the product improvement plan should be adopted: on the one hand, the focus is to improve the "fruit surface" attribute, that is, strictly control the fruit picking and fresh-keeping link, strictly control the screening link, improve the packaging and select reliable logistics companies;on the other hand, it is also necessary to reasonably allocate resources to other attribute improvement according to the income of attribute improvement.Through the summary of the work and results of the full text, the following research conclusions are drawn: first, the consumer satisfaction of e-commerce agricultural fruit products is low and the market performance is relatively poor, and the competitiveness of e-commerce fruit market is weak;second, consumer preferences and needs may change over time, and e-commerce agricultural product merchants should pay timely attention to consumer preferences and needs, and timely adjust the operation strategy and resource allocation; Third, the improvement of competitors' product attributes may weaken the competitiveness of their own products, so e-commerce agricultural products merchants should also constantly improve their products to avoid falling into the trap of "not to advance is to retreat".

      According to the conclusions of the study, different suggestions have been put forward for different subjects of e-commerce to help agriculture: for consumers, first, you can actively make a true evaluation of the goods, second, the Jingdong platform is a good choice for fresh products such as fruit and support for consumption to help agriculture, and third, the purchase of fruits and other fresh products should be clear about the return of the compensation situation. For e-commerce to help farmers fruit merchants, one is to strictly control product quality, the second is to improve their own services, the third is to reasonably allocate resources to invest, the fourth is to keep up with consumer demand, grasp the consumer trend. For the platform, one is to optimize the construction of the logistics system, the second is to improve packaging management, and the third is to improve the after-sales service system.