Original scientific paper

CLASSIFICATION WITH NEURAL NETWORKS IN MANAGERIAL ACCOUTING

Brano Markić ; Faculty of Economies, University of Mostar, Mostar, Bosnia and Herzegovina

Abstract

Managerial accounting is oriented to producing information to managers – that is, people inside an organization who control and direct processes and operations. Therefore has to be prepared a large number of analysis and reports comparing actual results with plans or reports with many indicators about business results. Very often task in managerial accounting is classification. A classification problems occur when an object must to be assigned into predefined group or class based on the values of observed attributes. The traditional approach to classification is Bayesian theory of decision making. In the paper is analysed another approach based on neural networks and their linking and integration with cluster analyse ( k-means algorithm). It was built CLUSTERNET system which from data warehouse according to manager’s requirements produces a wanted number of clusters and using neural network, built in language R, classifies each one new object into corresponding cluster.

Keywords

neural network; classification; managerial accounting; clustering

Hrčak ID:

72601

URI

https://hrcak.srce.hr/72601

Publication date:

26.9.2011.

Article data in other languages: croatian

Visits: 2.390 *