Showing 1–1 of 1 results for author: Smelkova, E

Generalization of metric classification algorithms for sequences classification and labelling

Authors: Roman Samarev, Andrey Vasnetsov, Elizaveta Smelkova

Abstract: The article deals with the issue of modification of metric classification algorithms. In particular, it studies the algorithm k-Nearest Neighbours for its application to sequential data. A method of generalization of metric classification algorithms is proposed. As a part of it, there has been developed an algorithm for solving the problem of classification and labelling of sequential data. The ad… ▽ More The article deals with the issue of modification of metric classification algorithms. In particular, it studies the algorithm k-Nearest Neighbours for its application to sequential data. A method of generalization of metric classification algorithms is proposed. As a part of it, there has been developed an algorithm for solving the problem of classification and labelling of sequential data. The advantages of the developed algorithm of classification in comparison with the existing one are also discussed in the article. There is a comparison of the effectiveness of the proposed algorithm with the algorithm of CRF in the task of chunking in the open data set CoNLL2000. △ Less

Submitted 16 December, 2016; v1 submitted 15 October, 2016; originally announced October 2016.