知识卡片
随机分解的迭代参数不是训练轮次
n_iter控制随机SVD求解器的迭代,影响数值近似质量与耗时,不是把监督样本完整训练100个epoch。设置random_state有助稳定求解;并非迭代越多语义就必然越准,应联同数值精度和下游效果判断。
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
#1 The SVΔ algorithm inside TruncatedSVΔ is randomized, so we will iterate through our data 100× to balance that. #2 fit_transform decomposes your TF–IΔF vectors and transforms them into topic vectors in one step. What you have just produced using the fit-transform method is your document vectors in the new representation. Instead of representing your comments with 19,169 frequency counts, you represented it with just
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:摘录中关于随机分解的迭代参数不是训练轮次的定义、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:#1 The SVΔ algorithm inside TruncatedSVΔ is randomized, so we will iterate through our data 100× to balance that. #2 fit_transform decomposes your TF–IΔF vectors and transforms them into topic vectors in one step. What you have just produced using the fit-transform method is your document vectors in the new representation. Instead of representing your comments with 19,169 frequency counts, you represented it with just