Development of preprocessing information algorithms for predictive models for IDSS in procurement management

Abstract

The article is devoted to the development of algorithms for preprocessing data and solving problems such as eliminating data incompleteness and converting raw information into a form convenient for analysis. When obtaining real time series data, we are faced with the fact that they only occasionally appear in the required format. A significant part of the raw data sets is poorly structured or contains a large number of missing values. This makes it impossible to process the data later by some methods. In this regard, the main stages of data preprocessing are implemented, such as data cleaning, filling in missing values, smoothing the time series, normalization of variables. Thus, a set of data was obtained for further research, implementation and testing for predictive models for IDSS in procurement management.