Development of algorithms for processing time series when working with statistical reporting forms of the production sector of the penitentiary system
Abstract
Development of algorithms for processing time series when working with statistical reporting forms of the production sector of the penitentiary system
Incoming article date: 09.01.2024To date, the penitentiary system of the Russian Federation has collected quite extensive databases for the production sector. The collected data is a time series. However, when studying the mutual distributions of parameters, a number of problems arise, the main one of which is that a different data accounting system is maintained for different parameters: in some cases, data accounting is cumulative throughout the year, in other cases, actual values are taken into account (in other words, some time series are trending, while others are seasonal (cyclical)). Data accounting periods also differ: monthly, quarterly, or per year. Thus, at first glance, the related parameters are almost impossible to compare. The paper proposes a number of algorithms that would solve this problem. The aim of the work was to develop new algorithms that allow comparing trend and seasonal time series using the example of the industrial sector of the penitentiary system. The objectives of the study can be designated as: classification of parameters that are taken into account as seasonal and as trend time series; development of algorithms for their comparison; study of the applicability of the results obtained.
Keywords: algorithm, data processing, python, time series, penitentiary system, manufacturing sector.