Table 2 Labeling approach for lag \(=\) 1 and lag \(=\) 2
From: Activity-based Twitter sampling for content-based and user-centric prediction models
Lag = 1 | Lag = 2 |
|---|---|
\(z_1 \rightarrow l_1 : sgn | y_2-y_1| = +1\) | \(z_1 \rightarrow l_1 : sgn | y_3-y_1| = +1\) |
\(z_2 \rightarrow l_2 : sgn |y_3-y_2| = +1\) | \(z_2 \rightarrow l_2 : sgn |y_4-y_2| = +1\) |
\(z_3 \rightarrow l_3: sgn| y_4-y_3| = +1\) | \(z_3 \rightarrow l_3: sgn| y_{5}-y_3| = -1\) |
\(z_4 \rightarrow l_4: sgn| y_5-y_4| = -1\) | \(z_4 \rightarrow l_4: sgn| y_{6}-y_4| = -1\) |
\(z_5 \rightarrow l_5: sgn| y_6-y_5| = -1\) | \(z_5 \rightarrow l_5: sgn| y_{7}-y_5| = -1\) |
\(z_6 \rightarrow l_6: sgn| y_7-y_6| = -1\) | \(z_6 \rightarrow l_6: sgn| y_{8}-y_6| = -1\) |
\(z_7 \rightarrow l_7: sgn| y_8-y_7| = +1\) | \(z_7 \rightarrow l_7: sgn| y_{9}-y_7| = +1\) |