Yes, my concern is we don't know for sure that Aleksander had anything to do with Scorecard. I don't want to spread a false rumor. We know he worked at Dominion, we know he went to significant lengths to hide that, including very recent actions, we know he is an expert with such topics as "SMOTE (Synthetic Minority Over-sampling Technique)" and other such technology. I am not technical enough to know what any of that means. More research is needed.
I agree. There is no DIRECT connection to scorecard. BUT the details of what he did do led to the ability to make adjustments based on the top data point to keep the bottom data point on track to balance, or Tie.
AND
Philip Kegelmeyer's SMOTE, what SMOTEboost is based off, was a pioneer in " machine learning and graph algorithms, especially as applied to ugly, obdurate, real-world data which is, perhaps deliberately, actively resistant to analysis."
These are classifiers where a distribution (think equation) is being altered for detecting things. Finding hidden patterns in data is pretty normal, especially when there’s not a lot of the kind you’re looking for.
Yes, my concern is we don't know for sure that Aleksander had anything to do with Scorecard. I don't want to spread a false rumor. We know he worked at Dominion, we know he went to significant lengths to hide that, including very recent actions, we know he is an expert with such topics as "SMOTE (Synthetic Minority Over-sampling Technique)" and other such technology. I am not technical enough to know what any of that means. More research is needed.
What he used was SMOTEboost. Here is the paper.
https://www3.nd.edu/~nchawla/papers/ECML03.pdf
I agree. There is no DIRECT connection to scorecard. BUT the details of what he did do led to the ability to make adjustments based on the top data point to keep the bottom data point on track to balance, or Tie.
AND
Philip Kegelmeyer's SMOTE, what SMOTEboost is based off, was a pioneer in " machine learning and graph algorithms, especially as applied to ugly, obdurate, real-world data which is, perhaps deliberately, actively resistant to analysis."
These are classifiers where a distribution (think equation) is being altered for detecting things. Finding hidden patterns in data is pretty normal, especially when there’s not a lot of the kind you’re looking for.