知识卡片
弱标签能提供样本但不是金标准
星级和情感标签可作监督信号,话题标签也可辅助构造样本,但评论可能对不同方面评价不一致或带讽刺。应检查标签映射与噪声,且去掉直接透露标签的输入线索;均值为零更不证明数据没有偏差。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 思维的词元
原书出处 · 1 处
- · 查看原书依据
The second approach, machine learning, relies on a labeled set of statements or documents to train a machine learning model to create those rules. A machine learning sentiment model is trained to process input text and output a numerical value for the sentiment you are trying to measure, like positivity, “spamminess,” or “trolliness.” For the machine learning approach, you need a lot of text data
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第02章《思维的词元》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-2.html)
- 结论依据:摘录中关于弱标签能提供样本但不是金标准的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。
- 原始内容:The second approach, machine learning, relies on a labeled set of statements or documents to train a machine learning model to create those rules. A machine learning sentiment model is trained to process input text and output a numerical value for the sentiment you are trying to measure, like positivity, “spamminess,” or “trolliness.” For the machine learning approach, you need a lot of text data