Table 4 ARIMA best parameter combinations over the validation set for each topic on two different platforms
From: Experimental evaluation of baselines for forecasting social media timeseries
Domain | Topic | (p, q, d) | |
|---|---|---|---|
YouTube | |||
Vz19 | arrests | (72, 1, 48) | (48, 2, 72) |
international/aid | (24, 2, 48) | (72, 1, 96) | |
maduro/dictator | (96, 2, 72) | (96, 1, 72) | |
military | (72, 1, 72) | (72, 1, 96) | |
other/chavez | (48, 2, 72) | (48, 1, 24) | |
CPEC | benefits/development/roads | (24, 1, 72) | (72, 2, 72) |
controversies/china/border | (24, 2, 72) | (48, 2, 72) | |
controversies/china/uighur | (24, 1, 48) | (24, 2, 48) | |
controversies/pakistan/baloch | (48, 2, 96) | (48, 2, 72) | |
leadership/sharif | (24, 2, 48) | (24, 1, 24) | |
BRIA | covid | (96, 1, 48) | (72, 2, 96) |
debt | (96, 1, 96) | (48, 1, 96) | |
infrastructure | (24, 1, 24) | (72, 1, 96) | |
mistreatment | (96, 1, 96) | (48, 1, 48) | |
travel | (24, 1, 48) | (24, 2, 72) | |