Objective:compare the habits and features of obese (BMI>25) and normal (BMI<25) individuals and express a method to ameliorate the life styles using a cross-sectional experiment.Methods:A total of 220 randomly s...Objective:compare the habits and features of obese (BMI>25) and normal (BMI<25) individuals and express a method to ameliorate the life styles using a cross-sectional experiment.Methods:A total of 220 randomly selected cases were divided into case group (n=110) and control group (n=100) according to the calculated BMI level.Samples with BMI>25 kg/m2 were assigned to the case (obsess) group and those with BMI ranging from 20 to 25 were assigned to control (normal) group.The Miller-Smith life style questionnaires consisting 20 questions each with 5 different answers were given to both groups.Data of the questionnaires were collected and analyzed using t-test and Chi-square with SPSS.Results:No significant differences were found among the two groups in terms of the mean age,gender,level of education,marital status,insurance,breakfast,lunch or dinner,fried meat,legumes,caffeinated beverages,the length of sleep during 24 h,cigarette smoking and losing job or spouse.However,in regards to use of vegetables,sausage,fried potatoes,enriched breads,low fat milk,low salt,candies and chocolates significant relations were found (P<0.05).Conclusion:The present study suggests one way to control obesity and prevent diseases is to ameliorate the life styles.There is a relation between health and stress and irregularity of meals,such as breakfast skipping,is associated with overweight and obesity in adolescence.展开更多
针对现有的卷积、循环模型预测滚动轴承剩余使用寿命(Residual Life,RL)精度低的问题,提出一种基于改进自注意力机制的RL预测模型。首先,针对Transformer模型中自注意力机制内存占用高、信号存在噪声信息的问题,在窗口自注意力机制(Wind...针对现有的卷积、循环模型预测滚动轴承剩余使用寿命(Residual Life,RL)精度低的问题,提出一种基于改进自注意力机制的RL预测模型。首先,针对Transformer模型中自注意力机制内存占用高、信号存在噪声信息的问题,在窗口自注意力机制(Window Based Multi-head Self-attention,W-MSA)的基础上,提出概率窗口自注意力机制(Probwindow Based Multi-head Self-attention,PW-MSA);然后,针对多头信息不匹配和缺少局部信息的问题,采用Talking Head方法实现多头信息融合,并在前馈神经网络层加入深度可分离卷积提取局部信息,从而提升模型的预测精度。采用PHM2012轴承数据集将改进前后的自注意力机制模型进行比较,并和现有的先进预测模型对比,结果表明,改进自注意力机制模型可使预测精度提升13.04%。展开更多
文摘Objective:compare the habits and features of obese (BMI>25) and normal (BMI<25) individuals and express a method to ameliorate the life styles using a cross-sectional experiment.Methods:A total of 220 randomly selected cases were divided into case group (n=110) and control group (n=100) according to the calculated BMI level.Samples with BMI>25 kg/m2 were assigned to the case (obsess) group and those with BMI ranging from 20 to 25 were assigned to control (normal) group.The Miller-Smith life style questionnaires consisting 20 questions each with 5 different answers were given to both groups.Data of the questionnaires were collected and analyzed using t-test and Chi-square with SPSS.Results:No significant differences were found among the two groups in terms of the mean age,gender,level of education,marital status,insurance,breakfast,lunch or dinner,fried meat,legumes,caffeinated beverages,the length of sleep during 24 h,cigarette smoking and losing job or spouse.However,in regards to use of vegetables,sausage,fried potatoes,enriched breads,low fat milk,low salt,candies and chocolates significant relations were found (P<0.05).Conclusion:The present study suggests one way to control obesity and prevent diseases is to ameliorate the life styles.There is a relation between health and stress and irregularity of meals,such as breakfast skipping,is associated with overweight and obesity in adolescence.
文摘针对现有的卷积、循环模型预测滚动轴承剩余使用寿命(Residual Life,RL)精度低的问题,提出一种基于改进自注意力机制的RL预测模型。首先,针对Transformer模型中自注意力机制内存占用高、信号存在噪声信息的问题,在窗口自注意力机制(Window Based Multi-head Self-attention,W-MSA)的基础上,提出概率窗口自注意力机制(Probwindow Based Multi-head Self-attention,PW-MSA);然后,针对多头信息不匹配和缺少局部信息的问题,采用Talking Head方法实现多头信息融合,并在前馈神经网络层加入深度可分离卷积提取局部信息,从而提升模型的预测精度。采用PHM2012轴承数据集将改进前后的自注意力机制模型进行比较,并和现有的先进预测模型对比,结果表明,改进自注意力机制模型可使预测精度提升13.04%。
