将机器学习算法和文本挖掘融入酱卤肉制品货架期预测中,基于对文献数据库中酱卤肉制品的货架期及其影响因素(包装方式、储藏方式、保鲜剂和二次杀菌)进行收集,构建原始数据集;通过比较多种编码方法(JamesStein、BaseNEncoder、TargetEnc...将机器学习算法和文本挖掘融入酱卤肉制品货架期预测中,基于对文献数据库中酱卤肉制品的货架期及其影响因素(包装方式、储藏方式、保鲜剂和二次杀菌)进行收集,构建原始数据集;通过比较多种编码方法(JamesStein、BaseNEncoder、TargetEncoder、OrdinalEncoder、PolynomialEncoder),选择效果较好的JamesStein编码作为分类型特征变量的编码方式。通过比较多种机器学习算法(包括随机森林算法、K最近邻算法、逻辑回归、XGboost和多层感知机分类器),结果显示最优模型为随机森林算法[其准确度为0.95、精确度为0.97、曲线下面积(area under curve,AUC)值为0.99,F1-score 0.91]。通过对酱牛肉和盐水鸭的实际样品测试分析,发现该模型在预测不同酱卤肉制品的货架期方面均具有较高的准确性。此外,该文从另一个角度验证储藏温度、包装方式、保鲜剂和二次杀菌等因素对酱卤肉制品货架期的显著影响。展开更多
In Japan, 18.1% of the population known as baby-boomers will become the late-stage elderly in 2025, thereby needing a foundation to support this change. The Japanese Ministry of Health, Labour and Welfare is promoting...In Japan, 18.1% of the population known as baby-boomers will become the late-stage elderly in 2025, thereby needing a foundation to support this change. The Japanese Ministry of Health, Labour and Welfare is promoting the development of a regional comprehensive system allowing the elderly to continue living in their familiar surroundings. However, a care shortage is inevitable unless elders are able to age in good health, regardless of the system’s level of enhancement. This study aims to review the literature on active aging, clarify trends in clinical operations undertakings and research in Japan, and consider relevant research issues. After combining the search results of “active aging” and “healthy life expectancy,” we used a text mining technique to analyze the abstracts of 120 original articles and 213 reviews, commentaries, and features. Eight categories were extracted from the original articles: health statistics, gender, age, etc. From the reviews, commentaries, and features, 16 categories were extracted: orientation, disease, and living, etc. Cerebrovascular disease and osteoporosis were the most common diseases covered in the original articles;there has been a substantial amount of research on “active aging” and “healthy life expectancy” because they can easily lead to being bedridden and to a decrease in QOL. In the reviews, commentaries, and features, lifestyle-related diseases and menopause rather than cerebrovascular disease and osteoporosis, were extracted. The categorical differences found in the original articles may be due to the possibility that Japanese researchers are publishing their research abroad rather than in Japan or they submit research on topics that are guaranteed to be published at home or abroad. Little research has been conducted using the terms, “active aging” and “healthy life expectancy,” evidenced by the small number of studies generated. Preparations for 2025 will require an increase in the number of studies from the perspective of “active aging” and “healthy life expectancy.”展开更多
文摘将机器学习算法和文本挖掘融入酱卤肉制品货架期预测中,基于对文献数据库中酱卤肉制品的货架期及其影响因素(包装方式、储藏方式、保鲜剂和二次杀菌)进行收集,构建原始数据集;通过比较多种编码方法(JamesStein、BaseNEncoder、TargetEncoder、OrdinalEncoder、PolynomialEncoder),选择效果较好的JamesStein编码作为分类型特征变量的编码方式。通过比较多种机器学习算法(包括随机森林算法、K最近邻算法、逻辑回归、XGboost和多层感知机分类器),结果显示最优模型为随机森林算法[其准确度为0.95、精确度为0.97、曲线下面积(area under curve,AUC)值为0.99,F1-score 0.91]。通过对酱牛肉和盐水鸭的实际样品测试分析,发现该模型在预测不同酱卤肉制品的货架期方面均具有较高的准确性。此外,该文从另一个角度验证储藏温度、包装方式、保鲜剂和二次杀菌等因素对酱卤肉制品货架期的显著影响。
文摘In Japan, 18.1% of the population known as baby-boomers will become the late-stage elderly in 2025, thereby needing a foundation to support this change. The Japanese Ministry of Health, Labour and Welfare is promoting the development of a regional comprehensive system allowing the elderly to continue living in their familiar surroundings. However, a care shortage is inevitable unless elders are able to age in good health, regardless of the system’s level of enhancement. This study aims to review the literature on active aging, clarify trends in clinical operations undertakings and research in Japan, and consider relevant research issues. After combining the search results of “active aging” and “healthy life expectancy,” we used a text mining technique to analyze the abstracts of 120 original articles and 213 reviews, commentaries, and features. Eight categories were extracted from the original articles: health statistics, gender, age, etc. From the reviews, commentaries, and features, 16 categories were extracted: orientation, disease, and living, etc. Cerebrovascular disease and osteoporosis were the most common diseases covered in the original articles;there has been a substantial amount of research on “active aging” and “healthy life expectancy” because they can easily lead to being bedridden and to a decrease in QOL. In the reviews, commentaries, and features, lifestyle-related diseases and menopause rather than cerebrovascular disease and osteoporosis, were extracted. The categorical differences found in the original articles may be due to the possibility that Japanese researchers are publishing their research abroad rather than in Japan or they submit research on topics that are guaranteed to be published at home or abroad. Little research has been conducted using the terms, “active aging” and “healthy life expectancy,” evidenced by the small number of studies generated. Preparations for 2025 will require an increase in the number of studies from the perspective of “active aging” and “healthy life expectancy.”
