| Due to consumer’s living standards continuous improvement and the awareness of the health of consumers continue to enhance, the expenditure and the budget for disease prevention is also increased. The investment for health becomes a trends, the health care industry is also entered rapid development stage. Currently, health care products market is in a change period of adjustment and innovation, depends on solely traditional marketing methods cannot open the market in a short period and now combination of popular e-commerce and traditional channels will be a better choice. The customer is always the most valuable aspects of e-commerce resources. In order to achieve optimal network marketing, we need to grasp the customer’s mental state better and predict customer behavior trends through in-depth study of emotional perception of network users.This article uses a variety of techniques to collect, process, classification data. Using Web crawler technology to collect customer data pages and using X-means that improvement of clustering algorithm of K-means to complete the cluster computing based on customers’ offline behavioral data (browse, purchase, evaluation) and real-time behavior data (mouse, keyboard), and divides the customers into different behavioral characteristics groups. On the basis of domestic and foreign scholars’research results and the reference of ECC and PAD emotion model, establishes the two-dimensional emotional model to judge the emotional changes and tendency of buying to explore the customers’ potential needs and interests based on off-line and real-time customer behavior data and emotional state. Discover and analyze their emotions rule to achieve a high matching accuracy, better timeliness and higher user satisfaction, meanwhile also design emotion survey questionnaire and compare the the results of the questionnaire survey and analysis of customer behavior data to corroborate the the validity and usefulness of emotion model.This article uses off-line and real-time customers’ behavioral characteristics data to determine real-time emotional state changes of clients based on the traditional network data analysis. On the basis of off-line and real-time data, establishing two-dimensional emotional space model and comprehensive data based on multi-source data to measure customers’emotional dimension, avoid partiality of the emotional state and trends measurement of customers. At the same time, this article also provide important reference for the massive complex emotional data analysis for the future.This article is divided into five parts. The first chapter focuses on the research background and significance of the topic, and then describes an overview of this literature and materials, and finally introduces the main research content and direction. The second chapter describes the customer classification and behavior patterns of health care of online customer and the basic theory of sentiment analysis to clarify the relationship of customer behavior and emotion based on customer classification study. The third chapter studies sentiment analysis techniques, including automatic data access technology, affective computing and sentiment analysis systems. Chapter IV realize sentiment analysis technology and questionnaire system, and then discusses data analysis process through practical cases implementation, matching the results of data analysis and emotional model, finally discusses the practical application significance of emotional characteristics based on emotional model matching result. The final chapter summarizes the contents of the full text, raised the future direction of development of the field and further specializing. |
PDF Full Text Request