基于BP神经网络对阳虚质之虚寒症的模型研究

Application of BP Neural Network in the Cold of Insufficiency Type of Constitution of Yang Asthenia

摘要: 目的：以平和质的人群为对照组，基于BP神经网络对阳虚质之虚寒证进行建模研究。方法：通过选取中国化妆品研究中心对皮肤状态相关指标的体质的实测数据，利用Matlab软件采用BP神经网络建立专家得分(因变量)与这些指标(自变量)的量化关系用以研究平和质和阳虚质的分类问题，进而说明阳虚质之虚寒证的指标特点。结果：利用BP神经网络模型对平和质和阳虚质进行分类的正确率均达到90%以上。结论：此模型相比于直接利用主观判定体质而言更为客观、省时，可与主观分类相辅相成；另外，利用此模型不仅可以帮助消费者根据自身体质选购化妆品，还可以帮助化妆品企业针对消费者的不同体质，开发更有针对性的产品。

Abstract: Objective: Regarding a group of people with constitution of yin-yang harmony as control group to explore cold of insufficiency type of constitution of yang asthenia by using BP neural network. Method: Volunteers’ skin data were strictly collected in Chinese cosmetic research center. After data collected, Matlab software was used to establish by BP neural network relationship between expert scoring (subject data) and indicators (object data). Result: The advantage of this method is quick analysis people’s physique type and the accuracy of the result that classifying constitution of yin-yang harmony and constitution of yang asthenia is more than 90%. Conclusion: The advantage of this model is more objective and time-saving which not only can help consumers to choose ap-propriate cosmetic but also help cosmetic companies’ development more targeted products for different constitutions consumers.