[Objective] The research aimed to analyze change characteristics and forecast factors of the fog in Beibei District of Chongqing from 1953 to 2010. [Method] By observation data of the fog in Beibei District from 1953 ...[Objective] The research aimed to analyze change characteristics and forecast factors of the fog in Beibei District of Chongqing from 1953 to 2010. [Method] By observation data of the fog in Beibei District from 1953 to 2010, interdeoadal, interannual, seasonal and monthly varia- tion characteristics of the fog days and formation-dispersion time of the fog were conducted statistical analysis. Meteorological conditions and fore- cast factors of the fog were also analyzed. [Result] Distribution of the fog days in Beibei District had obvious interdecadal characteristics. Fog days was at its maximum in the 1980s while minimum in the 1960s. Fog duration presented slow increase trend. Interannual characteristic of the fog days overall presented increase trend, and it had 9-year periodic oscillation characteristic. Fog mainly concentrated in autumn and winter. Fog was mainly formed at night （20：00 -08：00） and dispersed in the daytime （08：00 -13：00）. Meteorological conditions which affected heavy fog in Beibei District were water vapor and stratification, wind field, temperature, relative humidity and so on. [ Conclusion] The research provided theoretical basis for scientific predication and forecast of the fog in Beibei District.展开更多
Fuzzy sets theory cannot describe the neutrality degreeof data, which has largely limited the objectivity of fuzzy time seriesin uncertain data forecasting. With this regard, a multi-factor highorderintuitionistic fuz...Fuzzy sets theory cannot describe the neutrality degreeof data, which has largely limited the objectivity of fuzzy time seriesin uncertain data forecasting. With this regard, a multi-factor highorderintuitionistic fuzzy time series forecasting model is built. Inthe new model, a fuzzy clustering algorithm is used to get unequalintervals, and a more objective technique for ascertaining membershipand non-membership functions of the intuitionistic fuzzy setis proposed. On these bases, forecast rules based on multidimensionalintuitionistic fuzzy modus ponens inference are established.Finally, contrast experiments on the daily mean temperature ofBeijing are carried out, which show that the novel model has aclear advantage of improving the forecast accuracy.展开更多
需求的波动性与对预测精度的高要求使得销售预测成为学界与业界研究的重难点问题,销售预测的准确性极大影响着企业的生产及最终收益。本文针对饮品零售行业的销售预测问题进行了详细的研究,构建了引入天气因素的改进HyperGBM机器学习预...需求的波动性与对预测精度的高要求使得销售预测成为学界与业界研究的重难点问题,销售预测的准确性极大影响着企业的生产及最终收益。本文针对饮品零售行业的销售预测问题进行了详细的研究,构建了引入天气因素的改进HyperGBM机器学习预测模型,全面比较了传统预测方法(如ARIMA、SARIMA和Prophet)与HyperGBM在预测准确性上的差异,并分析了加入天气因素对HyperGBM预测效果的影响。由于南北方气候差异较大,本文的研究选择西安代表北方的天气特征,选择昆明代表南方季节性特征显著的天气,并根据历史销售数据对SKU进行分类,分析适用于不同类别SKU的销售预测方法。基于企业提供的72个饮品的SKU历史数据的研究表明,HyperGBM在63个(共71个)SKU上相较于传统预测方法的预测效果更好,RMSE指标平均提升22.9%。对于天气因素的进一步研究表明,将天气数据融合到HyperGBM后,预测的准确度较无天气因素的模型最高提升了31.6%。本文根据季节趋势分解法(seasonal-trend decomposition using loess,STL)和增广迪基-富勒测试(augmented dickey-fuller,ADF)两项检测结果,将SKU按周期性和稳定性的强弱均匀划分为四个类别:周期性强且稳定、周期性弱且稳定、周期性强但不稳定、周期性弱但不稳定。研究发现,不同类别的SKU适用于不同的预测方法,周期性强的SKU类别适合采用SARIMA预测方法,稳定性强的SKU类别适合采用HyperGBM机器学习算法。本文的结论可以为饮品零售行业的销售预测提供帮助,指导企业按照商品类别选择适用的销售预测方法。展开更多
文摘[Objective] The research aimed to analyze change characteristics and forecast factors of the fog in Beibei District of Chongqing from 1953 to 2010. [Method] By observation data of the fog in Beibei District from 1953 to 2010, interdeoadal, interannual, seasonal and monthly varia- tion characteristics of the fog days and formation-dispersion time of the fog were conducted statistical analysis. Meteorological conditions and fore- cast factors of the fog were also analyzed. [Result] Distribution of the fog days in Beibei District had obvious interdecadal characteristics. Fog days was at its maximum in the 1980s while minimum in the 1960s. Fog duration presented slow increase trend. Interannual characteristic of the fog days overall presented increase trend, and it had 9-year periodic oscillation characteristic. Fog mainly concentrated in autumn and winter. Fog was mainly formed at night （20：00 -08：00） and dispersed in the daytime （08：00 -13：00）. Meteorological conditions which affected heavy fog in Beibei District were water vapor and stratification, wind field, temperature, relative humidity and so on. [ Conclusion] The research provided theoretical basis for scientific predication and forecast of the fog in Beibei District.
基金supported by the National Natural Science Foundation of China(61309022)
文摘Fuzzy sets theory cannot describe the neutrality degreeof data, which has largely limited the objectivity of fuzzy time seriesin uncertain data forecasting. With this regard, a multi-factor highorderintuitionistic fuzzy time series forecasting model is built. Inthe new model, a fuzzy clustering algorithm is used to get unequalintervals, and a more objective technique for ascertaining membershipand non-membership functions of the intuitionistic fuzzy setis proposed. On these bases, forecast rules based on multidimensionalintuitionistic fuzzy modus ponens inference are established.Finally, contrast experiments on the daily mean temperature ofBeijing are carried out, which show that the novel model has aclear advantage of improving the forecast accuracy.
文摘需求的波动性与对预测精度的高要求使得销售预测成为学界与业界研究的重难点问题,销售预测的准确性极大影响着企业的生产及最终收益。本文针对饮品零售行业的销售预测问题进行了详细的研究,构建了引入天气因素的改进HyperGBM机器学习预测模型,全面比较了传统预测方法(如ARIMA、SARIMA和Prophet)与HyperGBM在预测准确性上的差异,并分析了加入天气因素对HyperGBM预测效果的影响。由于南北方气候差异较大,本文的研究选择西安代表北方的天气特征,选择昆明代表南方季节性特征显著的天气,并根据历史销售数据对SKU进行分类,分析适用于不同类别SKU的销售预测方法。基于企业提供的72个饮品的SKU历史数据的研究表明,HyperGBM在63个(共71个)SKU上相较于传统预测方法的预测效果更好,RMSE指标平均提升22.9%。对于天气因素的进一步研究表明,将天气数据融合到HyperGBM后,预测的准确度较无天气因素的模型最高提升了31.6%。本文根据季节趋势分解法(seasonal-trend decomposition using loess,STL)和增广迪基-富勒测试(augmented dickey-fuller,ADF)两项检测结果,将SKU按周期性和稳定性的强弱均匀划分为四个类别:周期性强且稳定、周期性弱且稳定、周期性强但不稳定、周期性弱但不稳定。研究发现,不同类别的SKU适用于不同的预测方法,周期性强的SKU类别适合采用SARIMA预测方法,稳定性强的SKU类别适合采用HyperGBM机器学习算法。本文的结论可以为饮品零售行业的销售预测提供帮助,指导企业按照商品类别选择适用的销售预测方法。
