this is for holding javascript data
Mark S. Brown added ctrfitdiscussion.tex
over 10 years ago
Commit id: f60752fc29add1e4550298f7a9c11576b9d6214c
deletions | additions
diff --git a/ctrfitdiscussion.tex b/ctrfitdiscussion.tex
new file mode 100644
index 0000000..2545bf1
--- /dev/null
+++ b/ctrfitdiscussion.tex
...
In table \ref{tab:ctrfitresults} we perform two polynomial fits; namely the zeroth and 1st order. We determine the quality of these fits using $\chi^2$\footnote{I'm tempted to replace these with correlation coefficients} where values close to unity are good fits. $\chi^2_\text{nofit}$ is the reduced chi-squared test for no relationship between the CTR and DOI. $\chi^2_\text{linearfit}$ is the reduced chi-squared test for a linear relationship between the CTR and DOI. In the first instance we see the values are close to unity. Furthermore the error in the intercept$_\text{nofit}$ is larger than the peak to peak (PTP) change in the CTR values, indicating that there is no relationship between the CTR and DOI in all four data sets. Expanding to linear fits the $\chi^2_\text{linearfit}$ values are, on the whole, lower than unity indicating a parameter over specification.