Closed pstalidis closed 3 years ago
The reason that was detected was that in those cases the trustworthiness score of publishers was equally very large. The values are >5000 and after multiplying with the weight (0.5) the value of the fusion ends up in the range of 2500.
Publisher scores have been fixed so the values are now in the normal range
I have added a script to recalculate ALL the fusion scores in the articles already ingested
scores have been fixed
In some cases (~2k out of ~15k) the fusion score produces values outside the normal range, i.e. values > 2500 when the range is 0-100.