Open llrs opened 7 years ago
You are welcome to contribute :)
I found that this method is already implemented in R in the ppiPre package in the function TCSSGeneSim, maybe it is worth to explain it to your users. However I didn't manage to make it work. I will contact the maintainer to know if I am doing something or if it is still working, otherwise I would borrow part of the code. Also there is the IntelliGOGeneSim function with another algorithm of GO similarities. I will not implement it from scratch :hammer_and_wrench: :smile:
The ppiPre
package copy many source code from GOSemSim
and pretend that was implemented by themselves when they published a paper.
See https://guangchuangyu.github.io/2014/11/proper-use-of-gosemsim/.
I personally reject any source code derived from ppiPre
.
In case I don't manage to re-implement the TCSS algorithm, here is another similarity taking into account the relationships between edges. Which I think will be easier to implement albeit it doesn't use information content. (I post this here as a reminder to myself)
A list of methods can be found here.
Can anyone help me in implementing TCSS for finding the semantic similarity .can anyone help me out of this or kindly send me a link from where i can get some help
@Raheemkhan007 I could help, but I haven't worked in this in a while.
I could look at it again, you can create a pull request and I would review it if you want. Or we can comment your troubles in a separated issue/branch or fork.
@llrs i am working on it but i cant implement it in python. if you having any link according to it kindly share it with me
@Raheemkhan007 (almost) all information I have is posted here. If you implement it in python I can help in the translation to R. (Or it could be imported by R)
I won't be able to contribute with more implementations so I close the issue
I found another semantic similarity, are you interested in having it in the package?
I could spend some time trying to implement it in R for your package, or simply to call their program from the package. Here I left the reference: