情绪API显示错误分数正数而非中性 [英] Sentiment API showing incorrect score Positive instead of Neutral
问题描述
我在推特上使用Sentiment Score API。我有很多推文是正面的（有些我同意），但是当很多其他人在他们似乎没有包含任何积极的话语时，他们都是积极的。我注意到其中一些
包括谢谢这个词我想知道这是否有影响？
我已经通过它们运行它们：https：//azure.microsoft.com/en- gb / support / community /我得到了相同的结果。对我来说，Text API并没有向我显示结果我很有信心。
我的问题是，你有什么地方可以实际检查为什么分数会出现吗？目前这只显示一个没有任何计算的分数
Debbie
您好，
以下是对如何从$ b $对文档执行情绪分析的一些见解b 论坛。该论坛由产品组直接处理，使他们能够更好地了解这些问题，并帮助他们更准确地培训模型。
- -------------------------------------------------- -------------------------------------------------- -----
如果您觉得这篇文章有用，请给它一个"有帮助的"如果他们有帮助，请记得将回复标记为答案。
Im using the Sentiment Score API on tweets. I have lots of tweets coming through that are positive (And some I agree with) but then lots of others coming through as positive when they dont seem to contain any positive words to me. Ive noticed some of them include the word Thanks so Im wondering if this has an influence?
Ive run them through https://azure.microsoft.com/en-gb/support/community/ and I get the same results. To me, the Text API isn't showing me results Im confident in.
My Question is, Is there any where that you can actually check why the score comes out as it does? At the moment this just shows a score with none of the calculations behind it
Debbie
Hello,
Here is some insight on how sentiment analysis is performed on a document from the documentation.
Sentiment analysis is performed on the entire document, as opposed to extracting sentiment for a particular entity in the text. In practice, there is a tendency for scoring accuracy to improve when documents contain one or two sentences rather than a large block of text. During an objectivity assessment phase, the model determines whether a document as a whole is objective or contains sentiment. A document that is mostly objective does not progress to the sentiment detection phrase, resulting in a .50 score, with no further processing. For documents continuing in the pipeline, the next phase generates a score above or below .50, depending on the degree of sentiment detected in the document.
Currently, the result of the score is based on a pretrained model with extensive body of text with sentiment associations. It does not display the analysis of how the score is arrived/determined for a particular sentence. Currently, it is not possible to provide your own training data. The model uses a combination of techniques during text analysis, including text processing, part-of-speech analysis, word placement, and word associations. For more information about the algorithm, see Introducing Text Analytics.
In your case I think the score of some of the tweets that aren't positive might have been influenced by keyword like "Thanks", I tried this on some of the documents using the API and it looks like the inclusion of this word changed the score as you mentioned.
Without "Thanks"
With Thanks:
I would recommend to post your feedback and analysis on UserVoice forum. This forum is directly addressed by the product group and gives them good visibility on such issues and helps them to train the model more accurately.
-----------------------------------------------------------------------------------------------------------
If you found this post helpful, please give it a "Helpful" vote.
Please remember to mark the replies as answers if they help.
这篇关于情绪API显示错误分数正数而非中性的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
