Hi all, just a very brief question. Is the Simpson index the same as the Simpson -Yule index?

I have read a lot of papers where they mention these and didn't know if they are the same.

Best wishes

Hi all, just a very brief question. Is the Simpson index the same as the Simpson -Yule index?

I have read a lot of papers where they mention these and didn't know if they are the same.

Best wishes

Hello again Martyn,

I've never heard of the Yule index before now, so I looked up the original citation (Yule 1944).

Lots of metrics were invented/described multiple times by different people in different fields, so the same math get's multiple labels. I found this paper comparing Simpson and Yule.

After randomly selecting an individual, Yule's index measures "the probability of selecting a second individual that belongs to the same species." This is exactly the same same as Simpson's Index (D). Note that Simpson's Reciprocal Index (1/D) a.k.a. 'inverse Simpson' is also common.

Colin

P.S. Check out these sweet Yule quotes from the 1944 paper:

All the tables... have one characteristic in common, vis. that words occurring only once are the most frequent.

There is no immediately obvious theoretical reason for this.

Pretty sure that's how log distributions work, lol

As the size of the sample is increased: ...

the percentage of nouns used only once will fall slowly. .. if only we take a long enough work, words used twice... may prove to be the most frequent.

Dr. Anne Chao is going to make good use of that fact 40 years later in her own metric (PDF)

If a calculating machine is available, and it is essential for anyone who is doing much statistical work,

yeah

1 Like