Construction of a Novel Production Develop Decision Model Based on Text Mined

摘要

When users choose a product, they will consider the emotional experience triggered by the product form. The Kansei engineering is considered to be the most reliable and useful method to deal with users' emotional needs. Therefore, in this study a hybrid method that combines text mining and Kansei engineering is proposed, which have integrated TF-IDF, SD, BPNN, and NSGA-Ⅱ methods to extract product shape design solutions that meet user multidimensional needs. The TF-IDF is applied to analyze Kansei image factors of the product of user's review so as to realize the mining of user needs from the perspective of user real online shopping evaluation. Then, the FA is applied to analyze representative Kansei need items. Furthermore, the BPNN is used to identify the relationship between design variables and user demands, so that the prediction model is constructed. The non-dominated sorting genetic algorithm-Ⅱ is used as the multi-objective evolutionary method to obtain the Pareto optimal solutions that meets the user's multidimensional needs. Taking electric bicycles as an example, the experimental results show that this proposed method can help designers to obtain the production solutions based on users' real Kansei needs.