CC BY
24
Секция «Актуальные на учные проблемы в мире (глазами молодьш исследователей)»
УДК 339.18
ПРОГНОЗИРОВАНИЕ ПРИБЫЛИ МАГАЗИНОВ ТОРГОВОЙ СЕТИ "ROSSMANN"
Л. М. Савченко, А. Г. Юзаева Научный руководитель - В. В. Кукарцев Руководитель по иностранному языку - С. Г. Эфа
Сибирский государственный аэрокосмический университет имени академика М. Ф. Решетнева
Российская Федерация, 660037, г. Красноярск, просп. им. газ. «Красноярский рабочий», 31
Данная статья посвящена вопросам по построению модели прогнозирования прибыли на определенный период времени. Также рассмотрены различные методы прогнозирования. Для исследования сети магазинов Rossmann были использованы ее статистические данные и предложен способ прогнозирования прибыли с помощью data mining.
Ключевые слова: Data mining, моделирование, прогнозирование прибыли.
PROFIT FORECASTING OF THE RETAIL STORES "ROSSMANN"
L. M. Savchenko, A. G. Juzaeva Scientific Supervisor - V. V. Kukartsev Foreign Language Supervisor - S. G. Efa
Reshetnev Siberian State Aerospace University 31, Krasnoyarsky Rabochy Av., Krasnoyarsk, 660037, Russian Federation
This article is devoted to the construction of the model predicting income for a certain period of time. Also various methods of forecasting are discussed. There are studies taken for the chain of stores "Rossmann". Statistical data and the method ofprofit forecasting using data mining have been studied.
Keywords: operating Data mining, modeling, profit forecasting.
Profit forecasting is one of the most important aspects in the life of any commercial organization. This problem has been studying in many countries all over the world for a long period of time, since the main goal of any enterprise is making a profit. And due to the fact that English is one of the most spoken languages in the world, the exchange of experience between two countries is carried out in it. In turn, knowledge and experience from abroad make a significant contribution to the development of national companies. As management of organizations, including forecasting of earnings is largely borrowed from other countries, many experts in this field should be able to exchange skills and share their researches in English.
The issue of forecasting earnings has been investigated for a long time. We have got great experience, which helps to get closer to solving this problem.
Unfortunately, there is no model that will predict accurately profit of organizations, especially in a market economy.
First of all, it is necessary to examine the organization, for which it will be built: its mission and vision statements, company goals and philosophy, its position in the market. And there are many other questions which we should analyze. Analytics departments of large firms are involved in study of all these issues, and statistics collection. A model with a certain precision profit forecasting of the company based on that in the future is built. But you can not create a model that will predict your profit for any length of time. So, every model would predict only estimated profit for a short period of time.
Nowadays methods of forecasting can be divided into three groups: traditional methods, techniques of marginal analysis and economic-mathematical methods.
Traditional methods are divided into direct payment assortment, enlarged and combined calculation. Methods of marginal analysis are divided into the calculation of the break-even point and the "Input - Output - profit" plan based on the effect of operating leverage and financial planning on the basis of marginal cost and marginal revenue. Economic and mathematical methods include mathematical extrapolation,
Актуальные проблемы авиации и космонавтики - 2016. Том 2
mathematical statistics and probability theory, economic and statistical models. All these models are based on the parameters such as VAT, the price for the products, profitability and many other economic indicators.
Today, the chain of stores "Rossmann" is embarrassed by this problem. "Rossmann" has more than 3,000 stores in seven European countries. This is a chain of stores selling consumer goods: cosmetics, household cleaning products, personal care products, environmentally-friendly products and more. Analysts of store collect statistics on the basis and it helps you to create a predictive model of profits for up to 6 weeks. Statistical data have two bases. Firstly, you need to know the number of stores, store model, range, distance to the nearest competitor, month and year of the opening of a competitor, in-store events, week and year when promotion campaign started, etc. These entries describe each store; they are static and will not change over time. The only thing that can be changed is the number of stores.
The second database contains dynamic information: store number, date of observation, day of the week (observation), opening or closing the store this day, if there are any promotional campaigns there, if there is a public holiday or school holidays, etc. These two bases are the training sample, which will help to develop a model. Also, there is a test database, which contains the same information as the second base. Test data will be used as a criterion for the suitability of the model. The closer it will be to a real profit forecast, the better the model will be.
Developed and proven methods of forecasting earnings do not use the values that the retail chain "Rossmann" gives. Now the task is not only to predict, but to create a new method or some combination of already known methods. It is necessary to find information in the data, some of this information may be truly critical, and some may be irrelevant.
And data mining will help to solve this kind of problems. Data mining is an intelligent data analysis. This technology helps to identify hidden connections to databases. It can also handle very large database.
Data mining can perform predictive and descriptive modeling. In the case of predicting profits predictive modeling will be used. It solves such problems as regression, segmentation, time series analysis. The objective of regression is to identify the relationship between the state of the facility and factors that influence it. Segmentation identifies groups of objects that show the same behavior, as well as studying their features. Time series analysis is that the evaluation index is compiled by identifying trends, the seasonality of the data source.
Thus, the task of forecasting earnings set by the retail chain "Rossmann" must be resolved by creating a predictive model that works on the basis of synthesis methods of economic forecasting methods and data mining.
References
1. Bocharov V. V. Korporativnye finansy (Corporate finance). SPb. : Piter, 2008. 272 s.
2. Forecast sales using store, promotion, and competitor data [Electronic resource] : mode of access: https://www.kaggle.com/c7rossmann-store-sales (accessed: 03.03.2016).
3. Data mining - intellektual'ny analiz dannykh [Electronic resource]: mode of access: http://www. olap.ru/basic/dm2.aspire (accessed: 03.03.2016).
© Савченко Л. М., Юзаева А. Г., 2016
