# Knowledge base: Warsaw University of Technology

Back

## Comparative analysis of sales prediction systems based on neural networks

### Tomasz Sławiński

#### Abstract

Following thesis is written by Tomasz Sławiński, student of Management and Production Engineering at Warsaw University of Technology, graduate of XXXIII Bilingual Nicolaus Copernicus High School in Warsaw. The choice regarding subject of this project was made following author’s lifelong interest in broadly defined Information Technologies. Supported by partly economic profile of course taken throughout the years of studies and consultations with thesis supervisor resulted in final form, which stands as: ‘Systems of Sales Prediction in Enterprise, Based on Artificial Neural Networks’. Paper was supervised by PhD with habilitation of Engineering Tadeusz A. Grzeszczyk, Assistant Professor at Warsaw University of Science’s Faculty of Management. His educational activity regarding management based subjects consists, among others, of over 100, internationally known, publications, articles and books, in both English and Polish. Purpose of following thesis is to conduct a research on subjects regarding prediction systems based on artificial neural networks. They will be analysed pointing strengths and weaknesses. Predicted results of this thesis contain findings on economic effects involved in implementation of solutions of this kind. Calculation example contain of carrying out a sales forecast concerning volume of gas sales in the United States, using data consisting of volume observed in years 1995 to 2014. Using Statistica software and data sheets available on http://catalog.data.gov/dataset website, resulted in two predictions with values for January and February of 2015, and January through December of 2014. Results were later compared to real data observed. First two chapters are based on literature listed in bibliography. First one consists of brief presentation of prediction issue, explanation of terms regarding types of forecasts and forecasting methods with indicants defining its quality and accuracy. Basic prediction models are described with emphasis on those concerning time-series, which is later used in forecast carried out by the author. Second chapter summarize variety of artificial neuron models with theirs construction and architecture of networks consisting of them. Advantages and disadvantages of neural-based systems were outlined with hints regarding building them purpose-oriented since the beginning. Third chapter is practical realization of thesis purpose, which is conducting an artificial neural network based forecast of sales volume and comparing results with real observed data.
Diploma type
Engineer's / Bachelor of Science
Diploma type
Engineer's thesis
Author
Tomasz Sławiński (FoM) Tomasz Sławiński,, Faculty of Management (FoM)
Title in Polish
Analiza porównawcza systemów prognozowania sprzedaży bazujących na sieciach neuronowych
Supervisor
Certifying unit
Faculty of Management (FoM)
Affiliation unit
Study subject / specialization
, Zarządzanie i Inżynieria Produkcji
Language
(pl) Polish
Status
Finished
Defense Date
16-06-2016
Issue date (year)
2016
Reviewers
Michał Krawczyński (FoM/BITD) Michał Krawczyński,, Business Information Technology Department (FoM/BITD)Faculty of Management (FoM) Tadeusz A. Grzeszczyk (FoM/BITD) Tadeusz A. Grzeszczyk,, Business Information Technology Department (FoM/BITD)Faculty of Management (FoM)
Keywords in Polish
sztuczne neurony, sieci neuronowe, szeregi czasowe, prognozowanie sprzedaży, metody prognozowania, Statistica, poziom sprzedaży
Keywords in English
neural networks, artificial neurons, sales prediction, Statistica, time-series, sales level, prediction methods
Abstract in Polish
File
• File: 1
praca_pI_WZ_Slawinski_Tomasz_238118.pdf
Request a WCAG compliant version
Local fields
Identyfikator pracy APD: 12498

Uniform Resource Identifier
urn:pw-repo:WUT440d8d360096429a8e510585f22ead13