diff options
Diffstat (limited to 'paper/sections')
| -rw-r--r-- | paper/sections/appendix.tex | 8 |
1 files changed, 4 insertions, 4 deletions
diff --git a/paper/sections/appendix.tex b/paper/sections/appendix.tex index 5a6a2c7..32e2775 100644 --- a/paper/sections/appendix.tex +++ b/paper/sections/appendix.tex @@ -134,7 +134,7 @@ Figure~\ref{fig:running_time_n_nodes}.} We include here a running time analysis of our algorithm. In Figure~\ref{fig:running_time_n_nodes}, we compared our algorithm to the -benchmarks for increasing values of the number of nodes. In +benchmark algorithms for increasing values of the number of nodes. In Figure~\ref{fig:running_time_n_cascades}, we compared our algorithm to the benchmarks for a fixed graph but for increasing number of observed cascades. @@ -143,9 +143,9 @@ Even though both the MLE algorithm and the algorithm we introduced are based on convex optimization, the MLE algorithm is faster. This is due to the overhead caused by the $\ell_1$-regularisation in~\eqref{eq:pre-mle}. -The dependency as the number of cascades increases is linear, as expected. The -slope is largest for our algorithm, which is against caused by the overhead -induced by the $\ell_1$-regularization. +The dependency of the running time on the number of cascades increases is +linear, as expected. The slope is largest for our algorithm, which is against +caused by the overhead induced by the $\ell_1$-regularization. |