| 杨韫仪,裴卉宁,李书航,娄紫涵,李苑漠.基于多项Logistic回归的座面材质预测模型[J].包装工程,2021,42(16):182-187. |
| 基于多项Logistic回归的座面材质预测模型 |
| Seat Surface Material Prediction Model Based on Multiple Logistic Regression |
| 投稿时间：2021-04-09 |
| DOI：10.19554/j.cnki.1001-3563.2021.16.025 |
| 中文关键词: 多项Logistic回归 座椅材质 人体压力分布 预测模型 |
| 英文关键词: multiple Logistic regression seat material body pressure distribution prediction model |
| 基金项目:河北省自然科学基金资助项目(G2021202008) |
|
| 摘要点击次数: |
| 全文下载次数: |
| 中文摘要: |
| 目的 为促进我国邮轮市场的进一步发展，提高乘客的乘坐舒适度，对不同体态特征的乘客进行座面材质的个性化推荐。方法 提出基于多项Logistic回归的邮轮座面材质预测模型方法。结果 选择PP棉、聚氨酯泡沫塑料、乳胶3类市场上常见材质作为实验样本，通过TACTILUS人体压力分布测量系统采集了60位实验对象的人体压力分布数据;计算出SPD值用以反应压力分布的均匀程度，从而确定对应的推荐座面材质;依此数据进行回归模型的监督学习，将预测结果与真实结果比较，该模型结果的混淆矩阵体现了较优的模型精度。结论 将多项Logistic回归与经典模型进行比较，利用多项Logistic回归建立的预测模型能够获得各类别座面材质的推荐概率，模型结果更加直观且预测正确率略高，是适配于不同座面材质压力分布参数预测问题的研究方法，并且可推广该模型用于剔除舒适度水平较低的材质，缩小座面材质的备选范围，在乘客选择座椅时起到辅助设计优化的作用，帮助提高乘客对乘坐舒适性的满意程度。 |
| 英文摘要: |
| In order to promote the further development of China’s cruise market and improve passenger comfort, personalized recommendations for seat materials are made for passengers with different physical characteristics. The paper proposes a prediction method of cruise seat surface material prediction based on multiple logistic regression. Three kinds of common materials on the market, such as PP cotton, polyurethane foam and latex, are selected as the experimental samples. The body pressure distribution data of 60 experimental subjects were collected through the TACTILUS body pressure distribution measurement system; the SPD value is calculated to reflect the uniformity of the pressure distribution to determine the corresponding recommended seat surface material; the supervised learning of the regression model is carried out based on this data, and the prediction results are compared with the real results. The confusion matrix of the model results reflects the better model accuracy.Comparing the multiple Logistic regression with the classic model, the prediction model established by the multiple Logistic regression can obtain the recommended probability of each type of seat surface material. The model results are more intuitive and the prediction accuracy is slightly higher, which is suitable for different seat surface materials. The research method of the pressure distribution parameter prediction problem, and the model can be generalized to remove materials with lower comfort levels, narrow the choice of seat surface materials, play a role in assisting design optimization when passengers choose seats, and help improve passenger satisfaction with ride comfort. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
