Development of Diagnostic Method and Algorithm for Skin Type Based on Consumer Language and Sentiment

As the global cosmetics market is steadily growing, Korea's cosmetics market is also growing. In spite of the growth of the cosmetics industry, most of the recommendation methods based on the information received by the consumers without clearly knowing their skin type are based on the recommendation methods of existing cosmetics. Also, it is not easy to find recommendation methods based on accurate and objective indicators of the consumer's skin type. As for the customized cosmetics that have received much attention in recent years, the recommendation method is based on the self-questionnaire of the consumer or the skin type of the skin which is known to the user. However, the conventional skin type survey is often complicated and requires expertise to proceed on its own. In this process, the actual type of skin and the type of skin it is recognizing may be different, and the deviation from the desired ideal skin may be larger. Therefore, it is necessary to develop research method and analysis algorithm for diagnosis of accurate skin type and language which is easy for consumers to understand. In this study, we are going to make a diagnosis about current skin condition quantitatively through analyzing the current skin problem, distress, cosmetic use purpose and pursuing emotional distinction based on consumer language and emotion. The results of this study can be used as a guide for making customized cosmetics according to consumers' skin type and skin condition. Furthermore, if we combine cosmetics-related information and consumer reactions on various webs and apps, we expect it to be used as an index for recommending customized cosmetics.