A Study on the Extraction of Customer Satisfaction Factors Based on the Customer Satisfaction Model Using Text Review and Preview

Through the era of big data, the predictive power for customers has greatly improved. This helps consumers efficiently search for satisfactory products within a huge amount of product information. Already in commerce, quantifiable customer's product ratings and textual customer reviews enable consumers to recognize the quality of the product through evaluations of customers who have already purchased it, even if not purchased it. However, ratings and customer reviews were averaging the overall assessment of other customers. Thus, this paper tried to explore ways to personalized information by further embodying customer reviews. In order to segment the quality of the product, the custom satisfaction model proposed by Kano is framed. Customer reviews were analyzed through Latent Dirichlet Allocation(LDA) to classify the quality of the products recognized by the customer. After that, it was digitized through sentiment analysis and entered into the customer satisfaction model. Through this, Commerce expected that it would be possible to develop a more specific strategy for the quality perceived by the customer.