Compute frequency of use of irregular engish verbs. Data set for "A list of verbs" iOS App.
The fequency word data comes from http://invokeit.wordpress.com/frequency-word-lists/
As a study help I've added the hint parameter based on the classification proposed in this article
Rafa Barberá
http://github.com/rbarbera
http://twitter.com/rbarbera
- unicodecsv for writing CSV files with UTF-8 encoded strings
- R to generate the graphs in this document and to analize the data
The 5 more frequently used verbs are:
Verb | Freq |
---|---|
be/am/are/is | 0.225649 |
get | 0.071703 |
do/does | 0.070281 |
have | 0.069897 |
come | 0.055208 |
As you can see the verb To Be is very predominant, perhaps to much. We can explore the distribution to see how are distributed all the sample
As we intend to use a scroller to select the number of verbs in function of frequency of use, this kind of stepped distribution isnt' very usefull. Usually in this cases we use a logaritmic one (normalized between 0.0 and 1.0)
This will be the data that we use in owr project. The adventage of using the frecuency data to select the number of verbs instead of only sort the data by frecuency and select the number of verbs in a linear way can be see with a pair of histograms
The number of verbs in each bin are constant. But natural languages are diferents. There are a small number of high frequency use, a small number of low frequency use and a lot of medium frequency. This can be see clearly on the log histogram