| Implementing the context tree weighting method for context recognition (2007) | |||||||||||||||
Abstract | |||||||||||||||
| The context tree weighting method (CTW) is a competent statistics–based universal data com-pression algorithm proposed by Willems et al. in 1995. Motivated by the superior data compression performance of CTW over Lempel–Ziv based algorithms, we thoroughly investigate in this work the usability of CTW to applications involving content recognition and classification by means of data compression. Given a test file that needs to be classified into one of several reference files, the reference file which leads to the best compression of the test file, when both files are appended, is selected as the most probable match. Moreover, based on its algorithmic structure, we propose the concept of context tree freezing to improve the implementation of CTW for content recognition. We demonstrate the superiority of our implementation of CTW with freezing by comparing its performance with a wide range of data compression algorithms for content recognition problems such as language recognition, authorship attribution, and even DNA data classification. I. | |||||||||||||||
Publication details | |||||||||||||||
| |||||||||||||||