Skip to main content
Figure 7 | EPJ Data Science

Figure 7

From: Early detection of promoted campaigns on social media

Figure 7

Distributions of KNN-DTW classifier scores. We use Kernel Density Estimation (KDE), a non-parametric smoothing method, to estimate the probability densities based on finite data samples. We also show the threshold values that separate the two classes yielding an optimal F1 score.

Back to article page
\