¥140.00元 Azure Machine Learning forecasting time series based on different ARIMA models

发布于:2018-06-15 来自:freelancer 截止交稿时间截止交稿时间 2018-06-15 关注:22 评论:0

I'm looking for someone who can help me with setting up a simple azure machine learning project for forecasting time series based on ARIMA models

The routine should enable to

1) Drag and Drop time series (in Excel format)

2) enable to manually set-up to two different ARIMA models. The manual set-up of the two should concern (1) the quantity of data that the ARIMA model is processing (the section of the time series that is used to do the predictions), (2) the time lag between the data that is used for training the ARIMA model) and the prediction

3) a third ARIMA model, which is doing forecasts based on the mean of “ARIMA model 1” and “ARIMA model 2”. It should be possible to manually weight the influence of “ARIMA model 1” and “ARIMA model 2” for the third ARIMA model

4) compare the accuracy of the three different ARIMA based on mean absolute percentage error (MAPE)

The routine should be built based on this description: [login to view URL]

(Build and deploy forecasting models with Azure Machine Learning)

The person should:

- Setup the azure machine learning project with the goals as described above

- be available for possible follow-up projects

Requirements:

- Examples of early project you have done on Azure Machine Learning

- Examples / proof of other projects on forecasting time series (with ARIMA)

- High feedback/recommendations

Note: the project should start with this simple proof of concept. After this a simulation project is required. In the following project it should be possible to simulate the accuracy of the three models if different features of the time series (volatility) and the relationship of the ARIMA models 1 and 2 (difference in information processing quantity, difference in time lag) change. The results of the simulations should be visualized

Looking forward to the biddings.

Thanks and best regards

Moritz

技能: Azure, 机器学习

查看更多: Analyse 2 time series and propose the best way to get best forecasting mean, machine learning azure, azure machine learning, machine learning pattern matching time series, java source code time series forecasting, time series forecasting critical path, time series forecasting, time series forecasting spss excel sas, neural networks time series forecasting matlab, web based data mining machine learning projects bids, forex time series forecasting, time series analysis project, software time series analysis hurst exponent, part time home based job housewife, part time home based work

免责声明:

任务易所有内容均为威客和外包行业网站提供或收集于互联网公开的信息,目的是给在网络上工作的威客和兼职人员收集更多的免费工作信息,以帮助更多的人自主就业。如果有内容触及您的权益,请给我们发邮件(kf@renwuyi.com)并附上具体网址和说明,核实后我们将立即删除!对免责声明的解释、修改及更新权均属于任务易所有。

你觉得这个任务肿么样?

我坚信,评论可以一针见血。
暂无评论
  • 评分:
    3.0分
  • Freelancer.com是世界上最大的自由职业、外包和众包市场,拥有许多用户和项目。我们将来自全球超过247个国家、地区的超过 9,471,792 雇主和 自由职业者联系在一起。通过我们的市场,雇主可以雇用自由职业者完成例如从软件开发、写作、数据输入和设计到 工厂、科学、销售、会计和法律服务的工作。

你可能也对这些任务感兴趣

  • 暂无数据

便捷搜索

价格
日内的任务