I need a histogram with a logarithmic X axis; that is, the bin size is logarithmic (say with ranges 0.1-1, 1-10, 10-100, etc.). I'm storing the histogram as a plain 1D array.
Given a value, to find the corresponding bin, code I've found online loops over the bins, calculating bin width and checking if x falls in it, else continuing. That's O(num_bins), which is awful. How do I calculate the bin/array index in O(1) as is done in the linear case? Surely there's a closed-form analytic formula.
Edited by Prune, 29 November 2013 - 07:56 PM.