2023112701

v2023112701

2023112101

v2023112101

2023111301

v2023111301

2023110601

v2023110601

2023103101

v2023103101

2023102301

Changes:

• Uses an adjusted 4/3's earth radius to simulate standard atmospheric refraction

Info:

{
  "edges": {
    "edges": 3447329,
    "keys": 185820
  },
  "nodes": 4645
}

2023101701

The primary change was to implement 'carryover' from the previous denylist. This means that any nodes/edges that did not have any data for them over the 2 week sampling period are carried forward. This prevents hotspots from simply being turned off for two weeks to escape the denylist and then being turned back on.

All other denylist parameters remain the same as the previous week, and no new classifiers are added.

2023100802

2023100802

2023081001

2023081001

2023100510

This denylist changes some of the classifer parameters:

• Terrain intersection threshold lowered to 500kmm from 1000kmm
• Witness count limit for witness reciprocity lowered from 100 to 50
• Distance insensitivity curve relaxed by 500km

These changes reflect increased confidence in the terrain and witness reciprocity classifiers, and handles a case where we were too strict on hotspots 5km or less apart.