A Abdilmanova, S Sainova. Comparative quality estimation of recognition algorithms. Proceedings of the 13th International Scientific Conference INFORMATION TECHNOLOGIES AND MANAGEMENT 2015, April 16-17, 2015, ISMA University, Riga, Latvia, ISSN 1691-2489, рр. 73-75.
Abstract: In this research were defined the best recognition algorithm, with aim to train system to recognize particular rocks. The data applied in the research was got from Inkai uranium deposits, Kazakhstan. The system was trained on 4 and on 8 boreholes using three machine learning algorithms: Neural Network, k-NN and Decision Tree. Learning algorithms were tested on 1 borehole. Using predicted data, we chose indicators to define the quality of recognition. This research may direct future research on machine learning, what could lead to replacement of experts by machines.
Keywords: machine learning, k-NN, Neural Network, Decision Tree, training, testing, recognition, logging data, borehole.
1 Introduction
This research was done with aim to define the best learning algorithm of machine learning, by estimation of the quality recognition and classification.
In this research was analyzed different machine learning algorithms by applying them into logging data from Inkai uranium deposits.