Lohsi coefficient
To check the possible influence of TIC on the ranking of sites, for the values of the popularity index in the search engine " Vershki Runet " introduced correction factors equal to the values of ln (CY) of the corresponding sites.
The results were twofold: on the one hand, this at first glance improved the issue, giving more weight to the well-known sites of Runet, on the other hand, the effect of “leaving” large portals for non-core requests, for example, at the request of the “president”, news agency rian.ru and lenta.ru, as having one of the highest values of CY.
If the CY value were fixed at some interval, say from 0 to e ^ 10, and was calculated within each topic independently (i.e., all the “most important” sites within each topic would have the same highest rating), then such a the coefficient would be more useful in rankings (at least for Runet Tops :).
The results were twofold: on the one hand, this at first glance improved the issue, giving more weight to the well-known sites of Runet, on the other hand, the effect of “leaving” large portals for non-core requests, for example, at the request of the “president”, news agency rian.ru and lenta.ru, as having one of the highest values of CY.
If the CY value were fixed at some interval, say from 0 to e ^ 10, and was calculated within each topic independently (i.e., all the “most important” sites within each topic would have the same highest rating), then such a the coefficient would be more useful in rankings (at least for Runet Tops :).
')
Source: https://habr.com/ru/post/1358/
All Articles