图 书 馆
基于信用卡交易数据的销售预测Forecasting sales based on card transactions data |
|
| 课程网址: | http://videolectures.net/sikdd2015_moraru_forecasting_sales/ |
| 主讲教师: | Alexandra Moraru |
| 开课单位: | 约瑟夫·斯特凡学院 |
| 开课时间: | 2015-10-21 |
| 课程语种: | 英语 |
| 中文简介: |  智慧城市是当今研究问题中的重要主题，在从经济到交通，健康和生活方式的许多领域都具有深远的影响。本文解决的问题是特定产品类别的销售预测。我们介绍了三种回归算法的结果，这些算法适用于实时数据，用于预测汽油的每小时累积销售量。进行三个短期间隔的预测，即未来1、4和8小时。为了确定获得最佳结果所需的历史数据量，还进行了一项研究。 |
| 课程简介: | Smart cities are an important topic in today’s research problems, with high impact in many domains from economy to transportation, health and living style. The problem addressed in this paper is that of sales forecasting for a specific category of products. We present the results of three regression algorithms, applies on real live data, for predicting the cumulative hourly sales of petrol. The prediction is made for three short term intervals, of 1, 4, and 8 hours into the future. A study has also been conducted in order to identify the amount of historical data required for optimal results. |
| 关 键 词: | 智慧城市; 回归算法; 销售预测 |
| 课程来源: | 视频讲座网 |
| 数据采集: | 2020-11-04：zyk |
| 最后编审: | 2021-01-08：yumf |
| 阅读次数: | 40 |