this is for holding javascript data
Lucas Fidon edited subsubsection_sparsity_problem_However_with__.tex
almost 8 years ago
Commit id: b38f35b9ec3746ed8bf37c239bce006974a735ef
deletions | additions
diff --git a/subsubsection_sparsity_problem_However_with__.tex b/subsubsection_sparsity_problem_However_with__.tex
index 8d2ff41..f15ecd5 100644
--- a/subsubsection_sparsity_problem_However_with__.tex
+++ b/subsubsection_sparsity_problem_However_with__.tex
...
\subsubsection{sparsity problem}
However with the histogram method the empirical distribution remain too sparse, which lead to inconsistant empirical mutual information values. Indeed the feature space of positions or accelerations are 2-dimensional
($p(x,y)$) \Big($p(x,y)$\Big) and 4-dimensional
($p(x_{1},y_{1},x_{2},y_{2})$) \Big($p(x_{1},y_{1},x_{2},y_{2})$\Big) whereas our data is in both case a 1-dimensional path in the feature space. So histogram of trajectories are liable to be sparse.
To cope with sparsity we used
Parzen windowing \textbf{Parzen windowing} as it is described in \cite{Pluim_2003} for instance.
Given a trajectory $T$, the probability $p(x,y)$ of $(x,y)$ is the sum of the contribution of each $(x',y')$ in $T$. The contributions are functions of a Gaussian kernel.
Hence the following definition of the probability of $(x,y)$ given $T$: