Closed mmiladi closed 7 years ago
Hi @fabriziocosta, Today @martin-mann and I were discussing about the benchmark. The conclusion was without having this locality feature, it would not be a fair comparison of the tools.
Ok guys, I'll work on it.
ok, done. the readme still needs to be updated. Note that EDeN has to be updated.
check the notebook for usage. now there is a new parameter: importance_semi_window e.g. print rase.stability_score(0,'A', importance_semi_window=5)
Wonderful @fabriziocosta ! Is it a rectangular weight function with a single width option, or a trapezoid one with two window width options?
in all cases the weights are given as a single rectangular window, inside the window the weight is 1, outside is 0 there are two ways to specify the window:
To get a (super) local comparison score between wild and mutant feature graphs: An option to compare and calculate RaSE score on a sub-graph surrounding the SNP location. This option can be implemented to set vector weight(?) of distant locations to zero.