文章摘要
董圣泽,王肖烨,王若羽,杨景浩,郭凌志.基于BP神经网络的用户感性评价模型构建及应用[J].包装工程,2025,46(2):82-90.
基于BP神经网络的用户感性评价模型构建及应用
Construction of Users' Emotional Evaluation Model Based on BP Neural Network and Product Modeling Design of Rice Cooker
投稿时间:2024-08-15  
DOI:10.19554/j.cnki.1001-3563.2025.02.008
中文关键词: 用户感性评价模型  BP神经网络  感性意象评价  电饭煲
英文关键词: user emotional evaluation model  BP neural network  emotional imagery evaluation  rice cooker
基金项目:陕西省自主技术创新引导计划项目(2023YFBT-22-01);宝鸡文理学院研究生创新科研重点项目(YJSCX23ZD09)
作者单位
董圣泽 宝鸡文理学院 机械工程学院陕西 宝鸡 721016 
王肖烨 宝鸡文理学院 机械工程学院陕西 宝鸡 721016 
王若羽 宝鸡文理学院 机械工程学院陕西 宝鸡 721016 
杨景浩 宝鸡文理学院 机械工程学院陕西 宝鸡 721016 
郭凌志 宝鸡文理学院 机械工程学院陕西 宝鸡 721016 
摘要点击次数:
全文下载次数:
中文摘要:
      目的 充分利用网络购物平台用户评论,寻找形态因子最佳组合以指导产品造型设计,解决部分产品难以契合用户感性需求的问题。方法 利用网络爬虫抓取某网络购物平台的用户评论并利用TF-IDF算法将其量化;使用主成分分析法,选取感性评价指标,借助形态分析法将目标产品分解为多个主要结构;运用BP神经网络构建用户感性评价模型,遍历所有形态因子组合以确定最优搭配。结果 以电饭煲为例,根据所构建模型可预测各评价指标最高的形态因子组合,该模型均方误差为0.004 9,决定系数为0.928 7,模型精度符合要求,利用问卷调查法进一步证明了预测结果有参考价值。结论 基于BP神经网络构建的模型拥有快速寻找最佳形态因子组合的能力,利用网络购物平台用户评论作为训练样本能够解决人工搜集或问卷调查获取样本时间长、成本高、市场响应慢、样本分布不均匀等问题。用户感性评价模型预测结果对设计师精准满足用户需求有重要的指导意义。
英文摘要:
      The work aims to fully leverage user reviews on online shopping platforms, identify the optimal combination of morphological factors for guiding product modeling design and address the challenge of aligning products with users' emotional needs. With rice cookers as an example, web crawling techniques were employed to collect user reviews from an online shopping platform and quantify them according to the TF-IDF algorithm. A principal component analysis was then conducted to select emotional evaluation indices, and the target product was decomposed into multiple main structures with the help of a morphological analysis. By constructing a BP neural network-based user emotional evaluation model and traversing all possible combinations of morphological factors, the highest-ranking factor combination for each evaluation index was determined according to the model. The mean square error of the model was 0.004 9 with a coefficient of determination of 0.928 7, indicating its high accuracy in meeting requirements. Furthermore, a questionnaire survey method validated that the prediction results held reference values. In conclusion, the constructed BP neural network model demonstrates its ability to efficiently determine the best combination of morphological factors. Utilizing user reviews from online shopping platforms as training samples overcomes challenges associated with time-consuming manual collection or questionnaire surveys such as high costs, slow market response times, and uneven sample distributions. Ultimately, the user emotional evaluation model's prediction results provide valuable guidance for designers seeking accurate product modeling schemes based on users' preferences.
查看全文   查看/发表评论  下载PDF阅读器
关闭

关于我们 | 联系我们 | 投诉建议 | 隐私保护

您是第23568734位访问者    渝ICP备15012534号-4

版权所有:《包装工程》编辑部 2014 All Rights Reserved

邮编:400039 电话:023—68792836传真:023—68792396 Email: designartj@126.com

    

  
 

渝公网安备 50010702501717号