this is for holding javascript data
Antonino Ingargiola Clarify the AIC and BIC criteria (2)
almost 8 years ago
Commit id: aa00b42f2243ce1f5dd0e34b07af01553f1dab8e
deletions | additions
diff --git a/FRETBursts authorea latex/full_article.tex b/FRETBursts authorea latex/full_article.tex
index a265d12..4bf7432 100644
--- a/FRETBursts authorea latex/full_article.tex
+++ b/FRETBursts authorea latex/full_article.tex
...
\end{lstlisting}
Other useful attributes are \verb|aic| and \verb|bic| which contain
respectively statistics for the Akaike information criterion (AIC)~\cite{akaike_new_1974}
and the Bayes Information criterion (BIC)~\cite{schwarz_estimating_1978}.
AIC and BIC are general-purpose statistical criteria for comparing
the
suitability of multiple non-nested models
and selecting according to the
most appropriate for a given dataset. data.
By penalizing models with higher number of parameters, these criteria
strike a balance between the need of achieving high goodness of fit
with the need of keeping the model complexity low to avoid overfitting.
diff --git a/fitting_2.tex b/fitting_2.tex
index 722cc9f..0d59751 100644
--- a/fitting_2.tex
+++ b/fitting_2.tex
...
\end{lstlisting}
Other useful attributes are \verb|aic| and \verb|bic| which contain
respectively statistics for the Akaike information criterion (AIC)~\cite{akaike_new_1974}
and the Bayes Information criterion (BIC)~\cite{schwarz_estimating_1978}.
AIC and BIC are general-purpose statistical criteria for comparing
the
suitability of multiple non-nested models
and selecting according to the
most appropriate for a given dataset. data.
By penalizing models with higher number of parameters, these criteria
strike a balance between the need of achieving high goodness of fit
with the need of keeping the model complexity low to avoid overfitting.