Geodesic
Neural network model for intervention analysis in the retail sector
Vestnik Sankt-Peterburgskogo universiteta. Prikladnaâ matematika, informatika, processy upravleniâ, no. 2 (2014), pp. 111-119 Cet article a éte moissonné depuis la source Math-Net.Ru

Voir la notice de l'article

This article proposes a differentiated method of assessing the impact of multiple simultaneous interventions that affect the performance of the company. The available auxiliary data about the objects of the company is used. The method is based on neural network algorithm for estimating the impact of interventions; herewith the standard architecture of the neural network is modified in that way which allows retrieving information about the impact of the occurred events for quantitative assessment of the effect of management decisions. For each intervention affecting the network performance, the function, which describes the nature of the influence of interventions, is proposed. The operation of this method is illustrated by the data of «Lenta» LLC. The proposed method includes a preliminary assessment of the influence of interventions that could have an impact on other objects of retail network. A comparison with the external evaluation of each intervention separately is held in the case of non-use of auxiliary data when constructing the ARIMA model, wherein the function of influence of an external event sought by selection method. It is shown that, despite the similarity of the aggregate result, avoiding the usage of auxiliary information leads to incorrect conclusions about the nature of the effect of each intervention separately. An evaluation of the long-term effect of external events and the total amount of losses in monetary terms is undertaken. Bibliogr. 9. Il. 6.
Keywords: neural networks
Mots-clés : ARIMA, intervention. 
@article{VSPUI_2014_2_a10,
     author = {N. A. Valiotti},
     title = {Neural network model for intervention analysis in the retail sector},
     journal = {Vestnik Sankt-Peterburgskogo universiteta. Prikladna\^a matematika, informatika, processy upravleni\^a},
     pages = {111--119},
     year = {2014},
     number = {2},
     language = {ru},
     url = {http://geodesic.mathdoc.fr/item/VSPUI_2014_2_a10/}
}
TY  - JOUR
AU  - N. A. Valiotti
TI  - Neural network model for intervention analysis in the retail sector
JO  - Vestnik Sankt-Peterburgskogo universiteta. Prikladnaâ matematika, informatika, processy upravleniâ
PY  - 2014
SP  - 111
EP  - 119
IS  - 2
UR  - http://geodesic.mathdoc.fr/item/VSPUI_2014_2_a10/
LA  - ru
ID  - VSPUI_2014_2_a10
ER  - 
%0 Journal Article
%A N. A. Valiotti
%T Neural network model for intervention analysis in the retail sector
%J Vestnik Sankt-Peterburgskogo universiteta. Prikladnaâ matematika, informatika, processy upravleniâ
%D 2014
%P 111-119
%N 2
%U http://geodesic.mathdoc.fr/item/VSPUI_2014_2_a10/
%G ru
%F VSPUI_2014_2_a10
N. A. Valiotti. Neural network model for intervention analysis in the retail sector. Vestnik Sankt-Peterburgskogo universiteta. Prikladnaâ matematika, informatika, processy upravleniâ, no. 2 (2014), pp. 111-119. http://geodesic.mathdoc.fr/item/VSPUI_2014_2_a10/

[1] Valiotti N., Abbakoumov V., “Using neural network models for intervention analysis”, J. Prikladnaya informatika, 2013, no. 5 (47), 6–13

[2] \href

[3] “Road sign on a Tallinnskoe highway turn was closed by piece of a cardboard”

[4] Bergstra J., Bengio Y., “Random Search for Hyper-Parameter Optimization”, Universite de Montreal. J. Machine Learning Research, 13 (2012), 281–305 | MR | Zbl

[5] Hyndman R., Khandkar Y., “Automatic Time Series Forecasting: The forecast Package for R”, J. of Statistical Software, 27:3 (2008), 8–19

[6] Haykin S., Neural Networks. A comprehensive foundation, Prentice-Hall, Inc., Toronto, Canada, 1999, 652 pp.

[7] Zhang G. P., Neural Networks in Business Forecasting, Idea Group Publishing, Georgia, USA, 2004, 350 pp.

[8] Hyndman R. J., Koehler A. B., Ord J. K., Snyder R. D., “Forecasting with Exponential Smoothing”, Softcover – The State Space Approach, 2008, 163–176

[9] Box G., Jenkins G., Time series analysis: Forecasting and control, Holden-Day, San Francisco, 1970, 46 pp. | MR | Zbl