知识卡片
词主题与文档主题矩阵应分清
若权重矩阵以主题为行、词项为列,每列就是一个词在全部主题上的坐标。文档列向量左乘该矩阵得到主题列向量;书中又将主题得分乘回词权重的标量表达,不应误叫同一个词的固定坐标。
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
- · 从词数到语义分析
原书出处 · 1 处
- · 查看原书依据
While reading these vectors, you might have realized that the relationships between words and topics can be “flipped.” The 3 × 6 matrix of three topic vectors can be transposed to produce topic weights for each word in your vocabulary. These vectors of weights would be the word vectors for your six words:
参考来源
- 位置:《Natural Language Processing in Action, Second Edition》第04章《从词数到语义分析》(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-4.html)
- 结论依据:摘录中关于词主题与文档主题矩阵应分清的定义、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:While reading these vectors, you might have realized that the relationships between words and topics can be “flipped.” The 3 × 6 matrix of three topic vectors can be transposed to produce topic weights for each word in your vocabulary. These vectors of weights would be the word vectors for your six words: