this can be radically improved by using what I am working on.
Using the distance as the metric for filtering (if distance > 2, filter) is suboptimal, as it limits discover ability.
What's better is a filter based on personalized Pagerank ( if PP < threshold, filter). PP looks both at the distance and at the "flux", meaning how many path exists thay connects you to another npub.
So, PP can be:
- more flexible, because you can choose the threshold in a continuoum
- improves the discovery (by for example allowing npubs at 3 hops if they are "high quality")
pippellia / Pip the social graph guy
npub176…7vgup
2024-10-10 20:45:13
in reply to nevent1q…37x8
Author Public Key
npub176p7sup477k5738qhxx0hk2n0cty2k5je5uvalzvkvwmw4tltmeqw7vgupPublished at
2024-10-10 20:45:13Event JSON
{
"id": "ad89be0ebc7af801019acf48d80d2c69afdb833318ead9f30a08edc682d6c87b",
"pubkey": "f683e87035f7ad4f44e0b98cfbd9537e16455a92cd38cefc4cb31db7557f5ef2",
"created_at": 1728593113,
"kind": 1,
"tags": [
[
"e",
"a82d1624f08739ffebf45f92dc26a1ebe6a25b6d8347b8b2999d097f628eba9e",
"",
"root"
],
[
"e",
"0a4aa3af51387a4c327c960c19bddce85d338758b4a26c4e298aaf4fa3e9fb78"
],
[
"e",
"ca0950b62e1c37101f929730bda4ab451d13a714153bbcb8a193b0234f53998a",
"",
"reply"
],
[
"p",
"0d06480b0c6e3be3c9a1a65d7e6bc2091227d55bf4c77eeb6037ba7776c300ec"
],
[
"p",
"4c800257a588a82849d049817c2bdaad984b25a45ad9f6dad66e47d3b47e3b2f"
]
],
"content": "this can be radically improved by using what I am working on.\n\nUsing the distance as the metric for filtering (if distance \u003e 2, filter) is suboptimal, as it limits discover ability.\n\nWhat's better is a filter based on personalized Pagerank ( if PP \u003c threshold, filter). PP looks both at the distance and at the \"flux\", meaning how many path exists thay connects you to another npub.\n\nSo, PP can be:\n- more flexible, because you can choose the threshold in a continuoum\n- improves the discovery (by for example allowing npubs at 3 hops if they are \"high quality\")",
"sig": "66a76208b3fb3a4d2a067c005c537db5c412e650f7706939932d71c8138d4685c11883715c976c493afc947be193fab6f336f0ec01d9261d2f9b2071ab065e9e"
}