npub1ky…jknq6 on Nostr: As explained in the GitHub, the process to recreate the model is as follows Step 0: ...
As explained in the GitHub, the process to recreate the model is as follows
Step 0: calculate the phase variable
step 1: calculate autoregressive (first lag) +phase multpile ols
i.e. Yh = a + b×Y(i-1) + c×phase
step 2: calculate decay func (in my case, I manually fit the damping function into the residuals in step 1)
step 3: calculate prediction from 1
step 4: add decay to step 3 - important! the order in which you add this matters due to the autoregressive nature. Follow the spreadsheet.
step 5: exponentiate the result
Published at
2023-02-21 06:25:34Event JSON
{
"id": "03e658bcacab78fdb445440180e6fb029b46a73b4e801e3d268667bce1e03df1",
"pubkey": "b101fdc9c6998ba2f786272ef85068d38dddf2d87853070622e1d6929091043a",
"created_at": 1676960734,
"kind": 1,
"tags": [
[
"e",
"f26d79fb6159e15428c7e7ab24d262191977ad7473d2a662d1effab198341da2"
],
[
"p",
"b101fdc9c6998ba2f786272ef85068d38dddf2d87853070622e1d6929091043a"
]
],
"content": "As explained in the GitHub, the process to recreate the model is as follows\nStep 0: calculate the phase variable \nstep 1: calculate autoregressive (first lag) +phase multpile ols \ni.e. Yh = a + b×Y(i-1) + c×phase\nstep 2: calculate decay func (in my case, I manually fit the damping function into the residuals in step 1)\nstep 3: calculate prediction from 1\nstep 4: add decay to step 3 - important! the order in which you add this matters due to the autoregressive nature. Follow the spreadsheet.\nstep 5: exponentiate the result",
"sig": "5829bae4590075149c3b357a531de780d4eef9e184a89730b36eb9628143956abb254c048c24593ca802cfc983bce748eee2b4edb8a6039a032458cd04f102c1"
}