diff options
author | Vasil Zlatanov <v@skozl.com> | 2018-12-14 11:52:02 +0000 |
---|---|---|
committer | Vasil Zlatanov <v@skozl.com> | 2018-12-14 11:52:02 +0000 |
commit | 42c214aedbfbda826c617e5081169dd8319b0c51 (patch) | |
tree | 29196cedf1752e62714e607497f31e3e94bce940 /report | |
parent | 1f174318840bfe4b2dcdafec0c7d74754afb6350 (diff) | |
download | vz215_np1915-42c214aedbfbda826c617e5081169dd8319b0c51.tar.gz vz215_np1915-42c214aedbfbda826c617e5081169dd8319b0c51.tar.bz2 vz215_np1915-42c214aedbfbda826c617e5081169dd8319b0c51.zip |
Remove spaces around lambda
Diffstat (limited to 'report')
-rw-r--r-- | report/paper.md | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/report/paper.md b/report/paper.md index a9a7388..7c57df9 100644 --- a/report/paper.md +++ b/report/paper.md @@ -199,7 +199,7 @@ improved rank-list: $d^*(p,g_i)=(1-\lambda)d_J(p,g_i)+\lambda d(p,g_i)$. The goal is to learn optimal values for $k_1,k_2$ and $\lambda$ using the training set, such that an improved Top-1 identification accuracy can be found. This is done using a multi-direction search algorithm to estimate $k_{1_{opt}}$ and $k_{2_{opt}}$ -and an exhaustive search for $\lambda$ from $ \lambda = 0 $ (exclusively Jaccard distance) to $ \lambda = 1 $ (only original distance) +and an exhaustive search for $\lambda$ from $\lambda = 0$ (exclusively Jaccard distance) to $\lambda = 1$ (only original distance) in steps of 0.1. The results obtained through this approach suggest: $k_{1_{opt}}=9, k_{2_{opt}}=3, 0.1\leq\lambda_{opt}\leq 0.3$. To verify optimisation of $k_{1_{opt}}$, $k_{2_{opt}}$ heat plots were performed heat on |