DATA PRESENTATION PROBLEM INVESTIGATION USING CORRELATION ANALYSIS IN RETAIL SALES MANAGEMENT PROCESS AUTOMATION TASK
One side of the problem of retail sales management process automating aspects, namely, the problem of presenting data of a large dimension is considered. The problem is that a large number of characteristics (parameters) are used to describe the data. One of the most popular approaches to reducing the dimension of feature space is correlation analysis are researched. The results of the correlation analysis application for data obtained as a result of the retail process through automated systems (vending machines) are presented. Parameters with strong dependencies were excluded by correlation analysis (in such parameters has a strong linear relationship). According to the calculated correlation correction, it was found that the detected strong linear relationship is not accidental. The results of the correlation analysis are confirmed by constructing a classifier based on an ensemble of decision trees forests. The classifier showed that there are no other strong dependencies in the parameters with a weak correlation connection. The results of the analysis were checked by an expert group (a group of marketing specialists from the retail industry), which confirmed their consistency.
Authors: N. V. N. V. Razmochaeva, D. M. Klionskiy .
Direction: Informatics, Computer Technologies And Control
Keywords: Data presentation, data analysis, data interpretation, correlation analysis, process automation, retail sales, management, machine learning
View full article