I used polynomial interpolation of degree 5 to approximate one 0 of the Bessel function $J_0$. I used the following data:
With the data above, I computed a polynomial using the images of the Bessel funcion as nodes and the $x$ values as the images. After that, I evaluated the polynomial at 0 to obtain the approximation of the 0.
I noticed that if I only use the 6 $x$ values with positive image I get a better approximation than if I use the ones with negative image, as well as a better approximation than if I use 3 with positive image and 3 with negative image (the closest ones to the 0 of the Bessel function). I cannot understand why this happens, any help? I hope I have explained myself clearly... I would guess this has something to do with the fact that the function is continuous and decreasingly monotone but I'm not sure... thank you in advance.