| Trust plays a decisive role in many social and economic activities with high uncertainty and risks,and is directly related to the final income of enterprises.The O2O medical beauty platform aims to create value for potential customers by providing real and effective information.However,the information asymmetry and the high risks of medical beauty services make it difficult for potential customers to generate trust,which is an important reason for the low conversion rate of medical beauty platforms.Online medical beauty service reviews is the most important reputation signal of sellers in the O2O medical beauty business.Usually,medical beauty service reviews can record the customer’s service experience both in text and image,aiming to improve potential customers’ trust in related products and services by disclosing the service quality.However,how to evaluate the impact of medical beauty service reviews on potential customers’ trust has not been studied in depth.In order to solve the information technology problem of predicting the trust degree of potential customers through medical beauty service reviews,this paper starts from two perspectives,which are behavioral science and design science.Firstly,from the perspective of behavioral science,a theoretical model was constructed to explain the influence of medical beauty service quality characteristic in reviews on potential customers’ trust under the theory of service quality and trust.Then,the theoretical model was validated by a linear regression model,which is the more service quality characteristic displayed in medical beauty service reviews,the higher the trust level of potential customers,and the attractiveness of reviewer positively moderates the relationship between service quality characteristic and trust level of potential customers.Specifically,the service quality characteristics shown in the medical beauty review of text have a positive impact on the three dimensions of trust including competence,integrity and benevolence,while the service quality characteristics shown in the medical beauty review of image only have a positive impact on the two dimensions of trust including competence and integrity.In addition,the attractiveness of reviewer has a positive moderating effect on the relationship between service quality characteristic shown in the medical beauty review of text and trust level of potential customers.Furthermore,from the perspective of design science,based on the empirical research results of the impact of medical beauty service review of text and image on potential customer trust,this paper constructs a deep learning model that integrates text and image characteristics of medical beauty service review to predict potential customer trust on medical beauty service.Compared with the classic machine learning model including support vector regression,random forest model and LightGBM model,the deep learning model has the strongest prediction ability.Finally,this paper uses a deep learning model based on feature fusion to predict trust level of new medical beauty service reviews that has not been manually coded.In the end,the predicted data verifies the relationship between medical beauty review characteristics and potential customer trust from the perspective of deep learning. |
PDF Full Text Request