--[ 2 - Similarities In the next two sections, we will present two algorithms which can be used with any set of elements. Elsim (included in the code archive at the end of this paper) is our implementation of those algorithms. It is open source software (LGPL), and with it you can compute similarities between any "described" elements :) ----[ 2.1 Between 2 sets of elements In this part we will describe how we can create a generic algorithm to search the similarities between 2 sets of elements. This algorithm can be used to compare all kind of elements if you are able to find a correct way to represent your data. For the comparison of our data, we will use the NCD, thus indirectly the Kolmogorov complexity. One of the major drawbacks [16] of the compression is the "time" required :) The tool can be powerful, but if you use a compression algorithm like LZMA, you are limited to "hello world" problems due to the speed of the compression. You need to choose carefully your (lossless) compressor. A compressor C is "normal" if the following properties (inequalities) are satisfied [17]: 1) Idempotency: C(xx) = C(x), and C(E) = 0, where E is the empty string. 2) Monotonicity: C(xy) >= C(x). 3) Symmetry: C(xy) = C(yx). 4) Distributivity: C(xy) + C(z)
Posted on: Tue, 18 Jun 2013 06:35:25 +0000
Trending Topics
Recently Viewed Topics
© 2015