From: Information cascades prediction with attention neural network
 | MSE (t = 1*) | MSE (t = 2*) | MSE (t = 3*) |
---|---|---|---|
 Features-linear (no \(\varvec{T}\)) | 4.106 | 3.823 | 3.715 |
 Features-linear (no \(\varvec{S}\)) | 3.976 | 3.640 | 3.524 |
 Features-linear | 3.821 | 3.511 | 3.423 |
 Proposed (no \(\varvec{T}\)) | 3.772 | 3.503 | 3.328 |
 Proposed (no \(\varvec{S}\)) | 3.716 | 3.540 | 3.407 |
 Proposed (time series \(\varvec{T}\)) | 3.809 | 3.621 | 3.463 |
 Proposed | 2.609 | 2.349 | 2.300 |
AMiner | |||
 Features-linear (no \(\varvec{T}\)) | 2.621 | 2.407 | 2.092 |
 Features-linear (no \(\varvec{S}\)) | 2.561 | 2.312 | 1.986 |
 Features-linear | 2.429 | 2.136 | 1.880 |
 Proposed (no \(\varvec{T}\)) | 2.411 | 2.050 | 1.799 |
 Proposed (no \(\varvec{S}\)) | 2.307 | 2.186 | 1.838 |
 Proposed (time series \(\varvec{T}\)) | 2.457 | 2.129 | 1.906 |
 Proposed | 2.172 | 1.672 | 1.534 |