|
|
Application statistical methods for estimation customer lifetime value
Samuseva, Katsiaryna ; Vraná, Lenka (advisor) ; Sobíšek, Lukáš (referee)
This bachelor thesis pay attention on statistical methods managing relationship with customers and different approaches of estimation customer lifetime value for the company in long time period. It is an empirical type of bachelor thesis, where we will test the theoretical assumptions on the real data. During this paper we research and measure customers behavior in order to obtain quantitative outputs to adaptation marketing strategy, efficient allocation of resources and choosing key clients. On the basis of past and present behavior we simulate expected in future level of purchasing behavior and thus try to determine the most profitable target groups for high-quality application of marketing tools. Main purposes of thesis are describing various attitudes to modeling CLV, applying the model Pareto/NBD on real date of the company Práce pro Vás, to evaluate forecast applicability in practice. Following tasks will be resolved in this paper: to introduce advantages and disadvantages of each modeling CLV approach, to evaluate quality of Pareto/NBD model and to determine the advanced extensions of the research in this issue. There are various attitudes to estimation of CLV, which are described in theoretical chapters, but practical part of this paper focuses primarily on application of probability model Pareto/NBD. With the help of Pareto/NBD managers can get the information about future customers activity, i.e. expected number of transactions from real behavior a probability that customer will be active in observation period. In nowadays reality of overloaded market, almost in every business sector, every company should rely on quantitative base of marketing and choose target clients, on which to focus their marketing efforts. Exactly this need matches with described and applied in this bachelor thesis model Pareto/NBD, with it help we will estimate customer lifetime value for strategic decisions and managing relationship with customers. Finally, we will resolve continuous tasks and will achieve all the determined goals.
|
|
|
Increasing the efficiency of advertising on Facebook by using decision trees
Randus, Jan ; Vraná, Lenka (advisor) ; Jašek, Pavel (referee)
This bachelor thesis aims to increase the efficiency of Facebook marketing campaigns, having used advanced statistical methods. For this purpose, I gathered data from Facebook marketing campaign of a fashion online store. In the theoretical part, the thesis addresses methods which were used for data analysis (decision trees) and also covers sampling methods which were used for model adjustments. In addition, the thesis describes the field of online performance marketing. The practical part is devoted to the application of decision trees and the creation of decision model on the data gathered in the beginning. Finally, the thesis answers the question how to advertise on Facebook as efficiently as possible, based on the output from my decision model.
|
|
|
Cross-channel attribution modelling
Žárský, Jiří ; Šperková, Lucie (advisor) ; Vraná, Lenka (referee)
This bachelor thesis focuses on attribution in the context of online marketing and studies available evaluation models for the performance of advertising campaigns. This performance is measured on the basis of the campaigns' effectiveness in catching the attention of customers and generating revenue. Data containing information about users' interactions with real advertising campaigns were used for the analysis. Prior to solving the attribution problem, data from the AdForm platform were cleansed and transformed into the required structures. This process is automated by the ETL tool called Keboola. Afterwards, data are analyzed using various attribution modelling techniques such as simple heuristics, the Shapley Value or Markov chains. The thesis discusses the theoretical side, as well as the actual application of these models. In the last section, the results of individual models are interpreted, taking into account the campaign costs. The interpretation is performed in the Tableau visualization software, using metrics such as the return on advertisement spending. This thesis presents a critical assessment of attribution models based on predetermined criteria. A scheme of data transformations, which can be used for future analyses of advertisement campaigns, was also created as part of this work. The thesis further includes a chapter discussing issues potentially leading to inaccuracies in the models' results.
|
|
|
DRG system
Vraná, Lenka ; Marek, Luboš (advisor) ; Mašek, Petr (referee)
This thesis is focused on description of the DRG (Diagnosis Related Groups) classification system and its application as the inpatient care financing tool. The objectives of this thesis are to sum up the possible applications of DRG system, to describe the calculation of indicators needed for health care payment in 2012 and to suggest some improvements to the algorithm to enhance the quality of data processing in the years ahead. The first part of this thesis is theoretical and it includes especially the history of the DRG classification system and an explanation of the basic concepts and the calculation methods. In the second part there is the description of the data file, which was used for the calculation of the inpatient care payments for 2012, its processing (specifically the determination of the relative weights of DRG) and the possibility to automate the whole solution.
|
|
|
Analysis of Temperature Time Series Measured by Stations Praha - Ruzyně and Brno - Tuřany between 1999 and 2008
Vraná, Lenka ; Helman, Karel (advisor) ; Bašta, Milan (referee)
This thesis is focused on analyzing time series of monthly average air temperatures measured at 7 AM, 2 PM and 9 PM, by climatological stations Brno - Tuřany and Praha - Ruzyně. The objective of this thesis is to employ basic time series analysis methods and to apply statistical software, concretely SPSS Statistics 17.0 and Statgraphics Centurion XVI. It deals especially with calculation of basic characteristics, time series decomposition (classical model) and description of their trend and seasonal component. This thesis examines if there is any resemblance in the progress of the series, which were measured by the same station at different times or at the same time by different stations. The thesis is divided into two parts. In the first part, there is a theoretical description of time series analysis methods and the analysis itself is in the second part.
|
guest ::
RSS feed.