提示
基于sda-moga算法的强关联商品补货量预测
首发时间：2024-07-09
摘要：不考虑商品关联销售的统计补货方案易造成销售短缺或库存积压。针对多销量情形的销售数据，提出一种对称分解的强关联规则挖掘算法，提取热销商品。以强关联规则支持度和置信度为目标，建立多目标遗传算法热销商品补货预测模型，设计一种目标差值计分方法构建适应度函数，并运用极限学习机预测补货量上限。选择uci销售数据集开展实验，结果表明提取至4-频繁项集的强关联规则置信度均在90%以上，建立的补货预测模型目标均值较最差的销量预测补货方案降低87.4%，较次优的销量均值补货方案降低74.9%。设计的适应度函数最优解目标均值较其它适应度函数降低50.1%至60.7%。
关键词：
for information in english, please click here
strongly associated commodity replenishment quantity prediction based on sda-moga algorithm
abstract：the statistical replenishment plan that does not consider the related sales of goods is easy to cause sales shortage or inventory overstocking.a strong association rule mining algorithm based on symmetric decomposition is proposed to extract hot-selling products from multi-sales data.aiming at the support degree and confidence degree of strong association rules,a multi-objective genetic algorithm replenishment prediction model of hot commodities is established,a goal difference scoring method is designed to construct fitness function,and an extreme learning machine is used to predict the upper limit of replenishment.the uci sales data set is selected to carry out the experiment.the results show that the confidence of the strong association rules extracted to the 4-frequent item set is above 90%,the target average value of the established replenishment prediction model is 87.4% lower than that of the worst sales prediction replenishment scheme,and 74.9% lower than that of the second-best sales average replenishment scheme.compared with other fitness functions, the average value of the optimal solution of the fitness function is reduced by 50.1% to 60.7%.
keywords：
论文图表：
引用
导出参考文献
no.****
动态公开评议
共计0人参与
勘误表
基于sda-moga算法的强关联商品补货量预测
评论
全部评论