The article deals with the problem of diagnosis of oncological diseases based on the analysis of DNA methylation data using algorithms of cluster analysis and supervised learning. The groups of genes are identified, methylation patterns of which significantly change when cancer appears. High accuracy is achieved in classification of patients impacted by different cancer types and in identification if the cell taken from a certain tissue is aberrant or normal. With method of cluster analysis two cancer types are highlighted for which the hypothesis was confirmed stating that among the people affected by certain cancer types there are groups with principally different methylation pattern.

Deep learning methods for image classification and object detection are overviewed. In particular we consider such deep models as autoencoders, restricted Boltzmann machines and convolutional neural networks. Existing software packages for deep learning problems are compared.

Software implementation of the gradient tree boosting algorithm is described and its competitiveness is illustrat-ed by the results of a computing experiment.

In this work the object detection problem is considered. A short description of implementations of the object detection system with a discriminatively trained part based model and a gradient boosting trees algorithm (as part of OpenCV library) is given. Application of the gradient boosting trees learner to the object detection problem (in terms of the pedestrian detection problem) is explored.

Павел Николаевич Дружков, кафедра математической логики и высшей алгебры, Нижегородский государственный университет им. Н.И.Лобачевского (Россия, г. Нижний Новгород), druzhkov_paul@mail.ru. Pavel Nikolaevich Druzhkov, Department of Mathematical Logic and Higher Algebra, N.I. Lobachevsky State University of Nizhni Novgorod (Russia, Nizhni Novgorod), druzhkov_paul@mail.ru. Николай Юрьевич Золотых, кандидат физико-математических наук, доцент, кафедра математической логики и высшей алгебры, Нижегородский государственный университет им. Н.И. Лобачевского (Россия, г. Нижний Новгород), nikolai.zolotykh@gmail.com. Nikolai Yur'evich Zolotykh, Candidate of Physico-mathematical Sciences, Department of Mathematical Logic and Higher Algebra, N.I. Lobachevsky State University of Nizhni Novgorod (Russia, Nizhni Novgorod), nikolai.zolotykh@gmail.com. Алексей Николаевич Половинкин, кафедра математического обеспечения ЭВМ, Нижегородский государственный университет им. Н.И. Лобачевского (Россия, г. Нижний Новгород), alexey.polovinkin@gmail.com. Aleksey Nikolayevich Polovinkin, Department of Software, N.I. Lobachevsky State University of Nizhni Novgorod (Russia, Nizhni Novgorod), alexey.polovinkin@gmail.com.

Several variations of parallel implementations of one of the supervised learning algorithms, Gradient Boosting Trees (GBT), with the use of Intel Threading Building Blocks are described. Results of experimental comparison and performance analysis of different approaches to parallelization are discussed.