TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Learning languages very quickly – with the help of some very basic Data Science

14 点作者 nns大约 8 年前

1 comment

x1798DE大约 8 年前
Quite misleading. This is just an explanation of how he assembled a word list to memorize - the top 1000 unique words used in two TV shows. Making word lists is not even close to the hardest thing about learning a language quickly, and 1000 words is not really a working vocabulary.<p>Additionally, this does not handle <i>most</i> of the things you&#x27;ll find annoying about assembling word lists. Imagine doing the same thing with English - you&#x27;ll probably find &quot;eat&quot;, &quot;eating&quot;, &quot;ate&quot;, &quot;eaten&quot;, &quot;buy&quot;, &quot;bought&quot;, &quot;buying&quot;, &quot;buys&quot; all in the top N &quot;words&quot;, but that&#x27;s hugely redundant information.<p>Another issue I&#x27;ve encountered when doing this sort of thing is that often you get words that have multiple meanings, and without actually understanding the context, you don&#x27;t know which one is the common meaning, and most common words are <i>hugely</i> overloaded - you can&#x27;t get a proper gestalt understanding of the concepts represented by a word just from reading the (often complicated) dictionary entry.<p>I&#x27;m very sympathetic to this kind of thing, since I&#x27;ve done similar things many, many times, but I think the author way over-sells what&#x27;s happening here, and is likely to run into many problems with this in the future.