From 50c719334f230d5b940868e5e128c96f7c088fcc Mon Sep 17 00:00:00 2001 From: nunzip Date: Wed, 12 Dec 2018 15:59:52 +0000 Subject: Evaluate rerank --- report2/paper.md | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) (limited to 'report2') diff --git a/report2/paper.md b/report2/paper.md index 8e76142..75989a1 100755 --- a/report2/paper.md +++ b/report2/paper.md @@ -50,7 +50,7 @@ be used as an alternative to euclidiean distance. To evaluate improvements brought by alternative distance learning metrics a baseline is established through nearest neighbour identification as previously described. Identification accuracies at top1, top5 and top10 are respectively 47%, 67% and 75% -(figure \ref{fig:baselineacc}). The mAP for a ranklist of size 10 is 33.3%. +(figure \ref{fig:baselineacc}). The mAP is 47%. \begin{figure} \begin{center} @@ -160,7 +160,15 @@ This is done through a simple **GRADIENT DESCENT** algorithm followed by exhaust $k_{1_{opt}}$ and $k_{2_{opt}}$ for eleven values of $\lambda$ from zero(only Jaccard distance) to one(only original distance) in steps of 0.1. The results obtained through this approach suggest: $k_{1_{opt}}=9, k_{2_{opt}}=3, 0.1\leq\lambda_{opt}\leq 0.3$. -## k-reciprocal Reranking Formulation +## k-reciprocal Reranking Evaluation + +Reranking achieves better results than the other baseline methods analyzed both as $top k$ +accuracy and mean average precision. +It is also necessary to estimate how precise the ranklist generated is. +For this reason an additional method of evaluation is introduced: mAP. + +It is possible to see in figure \ref{fig:ranklist2} how the ranklist generated for the same five queries of figure \ref{fig:eucrank} +has improved for the fifth query. The mAP improves from 47% to 61.7%. \begin{figure} \begin{center} @@ -170,6 +178,11 @@ in steps of 0.1. The results obtained through this approach suggest: $k_{1_{opt} \end{center} \end{figure} +Figure \ref{fig:compare} shows a comparison between $top k$ identification accuracies +obtained with the two methods. It is noticeable that the k-reciprocal reranking method significantly +improves the results even for top1, boosting identification accuracy from 47% to 56.5%. +The difference between the $top k$ accuracies of the two methods gets smaller as we increase k. + \begin{figure} \begin{center} \includegraphics[width=20em]{fig/comparison.pdf} -- cgit v1.2.3-54-g00ecf