Linux 软件免费装

在这里免费安装WordPress以及其他软件

Zipfstats

开发者 James Luberda
更新时间 2017年1月3日 10:14
PHP版本: 3.8.1 及以上
WordPress版本: 4.7
版权: GPLv2 or later
版权网址: 版权信息

标签

posts pages statistics analysis writing word count linguistics zipf. widget

下载

1.1 1.2

详情介绍:

(if you are not familiar with Zipf's law, you may wish to look at the screenshots first, as they provide a visual context for the following) Zipfstats is built around Zipf's law, which states that given some corpus (in this case, a post/page) of natural language, the frequency of any word is inversely proportional to its rank in the frequency table. In a perfect Zipfian distribution, this would mean that the most frequent word in the corpus will occur twice as often as the second most frequent word, three times as often as the third most frequent word and so on. Any given real world corpus, if large enough, will tend to approximate a Zipfian distribution (this also works for cities as ranked by population, though that observation is less useful in this context). The Zipfstats widget analyzes and reports word frequency by rank of a given single post or page (aka the corpus). The widget shows a small graph that plots the actual frequency-rank distribution of the content against a perfect Zipfian distribution, thus illustrating the degree to which the content follows Zipf's law. It also provides for a table of the top-ranked words. Zipfstats only analyzes the core content of a given post/page. It does not incorporate comments into its calculations. By default, it removes shortcode-generated content from its analysis, though you may optionally enable it. Filters, other than do_shortcode() (in the aforementioned case), are not affected in terms of the analysis. Actual output of the original content is unchanged, regardless of the options selected. Options Some things to note: Future Enhancements

安装:

  1. Upload zipfstats.zip to the /wp-content/plugins/ directory and extract the contents of the zipfile
  2. Activate the plugin through the 'Plugins' menu in WordPress
  3. Add the Zipfstats widget to a content bar

屏幕截图:

  • Widget screenshot showing default options.
  • The familiar "Hello world!" post with Zipfstats enabled. Note that each word appears exactly once, making it quite un-Zipfian, as the graph illustrates.
  • The familiar "Sample Page," which starts out fairly Zipfian, but as a significant number of words are used only once, ends in a very flat, un-Zipfian fashion.
  • Turning the GPLv3 license into a page with Zipfstats enabled, we can see how a much larger text corpus yields a plot much closer to a Zipfian distribution, as expected.
  • Finally, with a hand-crafted page designed to approximate a Zipfian distribution, we can see that the plot of actual words by frequency mostly fits the Zipfian distribution plot line (language has been abused to achieve this in a small corpus).

常见问题:

Why would I ever use this?

To find out how Zipfian your posts/pages are, and perhaps demonstrate that their content shows signs that the author is capable of using a natural language. Granted, Zipf's law is not generally regarded as prescriptive, meaning that it's not necessarily what a fluent writer in a given language should achieve, but there are arguments to the contrary. Mostly, because the graphs and tables are fun to look at (again, see screenshots).

Will it work with insert_theme_name_here?

In terms of calculations, Zipfstats only takes into account post/page content. Aspects of the widget display can be affected depending on theme configuration. Most Zipfstats work was performed against the Twenty Fourteen theme, as a point of reference.

Will it work with insert_plugin_name_here?

See above re: themes. If Zipfstats is set to strip shortcodes, then any content that is generated by a plugin's shortcode(s) will be ignored for purposes of analysis. Aside from that, plugin behavior (and, again, the content) will be unaffected.

Why do the numbers in the graph differ from those in the wordlist table/don't correspond to my manual word count?

It is traditional to plot the log of rank and frequency rather than the raw numbers when conducting a Zipf analysis. The table reflects the raw data (i.e. "the" may have appeared 237 times, but that figure is plotted as the log of 237). The key is that the degree of Zipfian fit is reflected in the graph (the more linear, the more Zipfian).

更新日志:

1.2 1.1 1.0