The procedure of building control charts with variable control limits based on nonparametric bootstrap percentile method is proposed. It makes possible to construct control chart of any statistic with unknown continuo...
The procedure of building control charts with variable control limits based on nonparametric bootstrap percentile method is proposed. It makes possible to construct control chart of any statistic with unknown continuous distribution and involves a large number of calculations. The order of building one of the scalar statistics to be monitored for the group of mutually dependent quality characteristics is considered. Formulas for obtaining variable control limits based on percentile method are shown. The issue of identifying the structure of relationships between quality characteristics is raised. Proposed control chart, Hotelling T2 and Shewhart usual control charts and their modifications for variable control limits have been constructed using actual data of two characteristics of the rolled steel manufacturing process. It is concluded that the development of the proposed approach can be useful in statistical process control.
The article describes the experience of teaching students of Big data analytics-related fields the basics of data science. The teaching methodology used at the Financial University under the Government of the Russian ...
The article describes the experience of teaching students of Big data analytics-related fields the basics of data science. The teaching methodology used at the Financial University under the Government of the Russian Federation is given, which consists in involvement of students of the group to competitions on Google Kaggle platform. In this methodology, the positive factors influencing the rapid acquisition of initial data science skills are: competitive effect; availability of prepared datasets and problem formulations from the real business sector in various directions; possibility of instant verification of the found solution; practical unlimited number of solutions. These factors contribute to the rapid understanding by students of the need to understand the essence of the problem for the best results, obligatory familiarization with the data, data cleaning and formatting, exploratory dataanalysis, forming a model with some basic level and its improvement, as well as interpretation.
A conceptual approach to the typology of countries' financial development has been developed, which will allow government authorities to make more accurate macroeconomic forecasts and apply a strategic approach to...
详细信息
It is well known that accurate probabilistic predictors can be trained through empirical risk minimisation with proper scoring rules as loss functions. While such learners capture so-called aleatoric uncertainty of pr...
详细信息
An assessment of the investor's risk profile is proposed as a risk coefficient in a model with a linear convolution of expected return and variance. The value of the risk coefficient is found from solving the opti...
详细信息
This article describes the author’s methodology for determining the genre of music from a melody using machinelearning and parallel data processing. For determination is used dataset formed by the author, on the bas...
This article describes the author’s methodology for determining the genre of music from a melody using machinelearning and parallel data processing. For determination is used dataset formed by the author, on the basis of which the model is trained and the genre is predicted on the test data. Yandex technologies, parallel programming, and Python data processing models are used in the work. There are also examples of training and test samples, model operation, data cleaning and dataset analysis methods. The work built neural network to identify genres and output results. The results vary from 35% to 50% and it should be noted that the dataset is rather small, so with bigger dataset we can await much higher result.
Large data is challenging for most existing discovery algorithms, for several reasons. First of all, such data leads to enormous hypothesis spaces, making exhaustive search infeasible. Second, many variants of essenti...
详细信息
Uncertainty quantification has received increasing attention in machinelearning in the recent past. In particular, a distinction between aleatoric and epistemic uncertainty has been found useful in this regard. The l...
详细信息
We consider the problem of estimating brain effective connectivity from electroencephalographic (EEG) measurements, which is challenging due to instantaneous correlations in the sensor data caused by volume conduction...
详细信息
Recently, a novel subspace decomposition method, termed 'Stationary Subspace analysis' (SSA), has been proposed by Bünau et al. [10]. SSA aims to find a linear projection to a lower dimensional subspace s...
详细信息
暂无评论