简体繁体 English

有关 Azure 文本分析中情绪分析频率的最佳实践

[英]Best practice about the frequency of sentiment analysis in Azure Text Analytics

原文 2019-07-02 20:02:08 2 1 asp.net/ azure/ azure-machine-learning-studio/ azure-analytics

In my asp.net application, I want to do a sentiment analysis on each discussion forum item as they are posted by the users.在我的 asp.net 应用程序中，我想对用户发布的每个论坛项目进行情绪分析。 I wonder if it is a good practice to make a request to Azure Text Analytics server to do a new Sentiment Analysis each time a text is posted by any user.我想知道每次任何用户发布文本时都向 Azure 文本分析服务器发出请求以进行新的情绪分析是否是一个好习惯。 Or, is it better to do this somehow once a day on all posts as a batch.或者，以某种方式每天对所有帖子批量执行此操作是否更好。 I wonder what is the best practice about this.我想知道这方面的最佳做法是什么。

1 个解决方案

We are working on a similar project as you.我们正在从事与您类似的项目。 My suggestion for your question is, you should select proper interval based on your requirement and need.我对您的问题的建议是，您应该根据您的要求和需要选择适当的时间间隔。 For example, if you want to react based on the sentiment result of every post in a timely manner, you need to do a analysis every time you pull a new post.例如，如果您想根据每个帖子的情感结果及时做出反应，则需要在每次拉新帖子时进行分析。 Also you can adjust your time interval of "pull and analyze" based on your business/ research need.您也可以根据您的业务/研究需要调整“拉取和分析”的时间间隔。 IF you just want to train your model and predict something based on the data you get and you have no need for timely reaction, I think once a day is enough.如果你只是想训练你的模型并根据你得到的数据预测一些事情，而你不需要及时的反应，我认为一天一次就足够了。

For our project, we are in the situation 1. So we will do a quick analysis when we receive any new post so that we can have a quick reaction.对于我们的项目，我们处于情况1。所以我们会在收到任何新帖子时进行快速分析，以便我们能够快速做出反应。

Hope this helps.希望这可以帮助。