Bonjour à tous,
Dans un papier traitant de computations neuronales au cours de décisions perceptives il est écrit:
The neuron's output in the presence and abscence of light can be characterized with a probability density function (PDF) which we will refer to as
f(e/h1) and f(e/h2) respectively. Given a particular output of this neuron (ek), it is possible to decide whether light is likely present (h1) or absent (h2) by computing the LR from the two PDFs. The likelihood of obtaining ek in the presence of light is the height of the ‘light present’ PDF, corresponding to f(e|h1), at the value ek. Similarly, the likelihood of obtaining ek in the absence of light is the height of the PDF corresponding to f(e|h2), at the value ek. A decision can then be made by determining whether the ratio of these likelihoods is greater or less than 1.
Pourquoi prend-on la hauteur de la fonction de densité pour déterminer les likelihoods et calculer ensuite le likelihood ratio ?
Help me please!
M.
-----