查询结果:   何海洪,余军合,许立波,李兴森.基于可拓聚类的服装需求预测方法[J].计算机应用与软件,2018,35(5):273 - 281.
中文标题
基于可拓聚类的服装需求预测方法
发表栏目
算法
摘要点击数
662
英文标题
CLOTHING DEMAND FORECASTING METHOD BASED ON EXTENSION CLUSTERING
作 者
何海洪 余军合 许立波 李兴森 He Haihong Yu Junhe Xu Libo Li Xingsen
作者单位
宁波大学机械工程与力学学院 浙江 宁波 315211 浙江大学宁波理工学院计算机与数据工程学院 浙江 宁波 315100    
英文单位
School of Mechanical Engineering and Mechanics, Ningbo University, Ningbo 315211, Zhejiang, China School of Computer and Data Engineering, Ningbo Institute of Technology, Zhejiang University, Ningbo 315100, Zhejiang, China    
关键词
服装业 需求预测 可拓聚类 K近邻 时间复杂度
Keywords
Clothing industry Demand forecasting Extension clustering K-nearest neighbor Time complexity
基金项目
国家自然科学基金项目(71271191,71671097);浙江省自然科学基金项目(LY16G010004,LY16G010010,LY18F020001);宁波市创新团队项目(2016C11024)
作者资料
〖HTH〗何海洪,硕士生,主研领域:数据挖掘,供应链管理。〖HTH〗余军合,副教授。〖HTH〗许立波,讲师。〖HTH〗李兴森,教授。 。
文章摘要
服装需求易受价格、消费习惯及气温、天气状况等随机因素影响而变得复杂多变难以预测,而实际服装需求预测中,区间预测往往比具体值预测更具指导意义。因此在分析价格、折扣、气温、天气状况等因素与服装销量相关性的基础上,采用可拓聚类方法构建服装需求区间预测模型,对服装需求进行区间预测,并给出具体的预测步骤与计算方法。结合实例数据,验证可拓聚类对服装需求区间预测的有效性,并与K近邻算法进行预测效果及时间复杂度对比分析。实例分析结果表明可拓聚类有比K近邻更好的预测效果,而时间复杂度却大幅缩减,此外可拓聚类还能体现预测值与各区间等级的隶属程度和亲疏关系。实例分析结果验证了可拓聚类应用于服装需求预测的有效性与优越性。
Abstract
Demand for garments can be complicated and unpredictable due to factors such as price, consumption habits, and temperature and weather conditions. In actual clothing demand forecasting, interval forecasting tends to be more instructive than specific value forecasting. Therefore, based on the analysis of the correlation between price, discount, temperature, weather conditions and clothing sales, the paper used the extension clustering method to build a clothing demand interval forecasting model. The clothing demand was forecasted in intervals, and the specific forecasting steps and calculation methods were given. Combined with the instance data, the effectiveness of the extension clustering on clothing demand interval prediction was verified. The K-nearest neighbor algorithm was used to compare the forecasting effect and the time complexity. The example analysis results showed that the extension clustering had better prediction effect than the K-nearest neighbor, but the time complexity was greatly reduced, in addition, the extension clustering also reflected the degree of membership and intimacy between the predicted value and each interval level. Which validated the effectiveness and superiority of extension clustering applied to clothing demand forecasting.
下载PDF全文