Predictive Modeling in selected database

Suada Myftari

Master's thesis

Abstract:

The area of technology that we are undergoing made Big Data and all the issues related to it, the topic of the day. Today companies are exposed to a huge flow of information but the key of success is how smart they use the huge information they own to improve their product and services based on the feedback of customers, how they improve the working condition of their employees to have satisfied and motivated people to want to work on that company and do their job well, how fast are they to make the changes and how efficiently they are in delivering their own product and services.This paper shows how Big Data creates value for the company, if used properly, through transparency, enabling analysis to discover what the needs are and to improve performance, segmentation to customize and personalize the strategies for every segment and what are the issues coming with Big Data as well as what are some of the techniques used for analyzing Big Data like cluster analysis, data mining, pattern recognition. Every process of business inside and outside the company is rich in information and therefore is a source of information. If companies pay attention to the data coming in and out and translating them in proper information and knowledge and to make smart use of them by making the right changes that they indicate, hence smarter and better decision and strategic business moves, they are already a step ahead and closer to satisfied customer which is key of their success and ongoing activity. The issues arising is what is considered Big Data and the Vs related to it and how they can be analyzed and processed and the case study presented in this research treats data mining and predictive modeling techniques to analyze categorical data in order to understand customer behavior and factors affecting this behavior. Decision trees(random forest, boosted trees and C&RT) are used to analyze the behavior and to see which among them has the higher accuracy in predicting factors affecting behavior of customers.

Keywords

Big data, V-s of big data, C&RT, Data preparation, Data cleaning, Random forest, Boosted trees, Data mining, Predictive modelling

Language used: English

Date on which the thesis was submitted / produced: 30. 11. 2016

Thesis defence

• Supervisor: Ing. Tomáš Hlavsa, Ph.D.

Czech University of Life Sciences Prague

Faculty of Economics and Management

Master programme / field:
Systems Engineering and Informatics / Informatics