DC FieldValueLanguage
dc.contributor.advisorMerkl, Wolfdieter-
dc.contributor.authorZafirovski, David-
dc.date.accessioned2021-04-20T12:54:38Z-
dc.date.issued2021-
dc.date.submitted2021-04-
dc.identifier.urihttps://doi.org/10.34726/hss.2021.78000-
dc.identifier.urihttp://hdl.handle.net/20.500.12708/17252-
dc.descriptionArbeit an der Bibliothek noch nicht eingelangt - Daten nicht geprüft-
dc.descriptionAbweichender Titel nach Übersetzung der Verfasserin/des Verfassers-
dc.description.abstractThe objective of the master thesis is to build a classification model using customer data, that classifies which customers will accept and which customers will reject a marketing campaign. The results of the predictive model can be used to segment and target correctly the customers who will respond positive and accept to the marketing campaign and to avoid targeting the customers who will respond negative and reject the marketing offer. This leads to an efficient work of the company, reduced costs and maximization of the company’s profit. The classification model was constructed using Cross-Industry Standard Process for Data Mining (CRISP-DM) methodology. A dataset about customer behavior and purchasing habits was collected from Kaggle data science online community. We have decided to choose this dataset because it covers our problem domain about personalized marketing and contains features from different types that describe personalized customer behavior about responses to past marketing campaigns, amount of money spent on different types of products, amount of deals conducted via various channels and personal customer data (year of birth, type of education, client’s registration date in the firm, amount visits to the firm’s web page in the past month, amount of days from the previous client’s purchase, material status, number of children, income). The target variable represents the customer’s response to the marketing campaign, whether the customer accepted or rejected the proposal in the latest marketing campaign. The research questions are, which of the features are the most important in the model construction for predicting the customer’s response to the marketing campaign based on personalized customer experience and which type of the selected classification algorithms provides the most precise prediction of customer’s response to the marketing campaign based on personalized customer experience. The machine learning algorithms that were used are: Logistic Regression, Random Forest, Support Vector Machine, Naïve Bayes and Multi-level Perceptron. The results showed that Logistic Regression classifier has the highest average Accuracy 0.87, highest average Precision 0.61, highest average F1 score 0.57 and highest average ROC AUC score 0.88. It was concluded that logistic regression provides the most precise prediction of customer’s response to the marketing campaign.en
dc.format65 Seiten-
dc.languageEnglish-
dc.language.isoen-
dc.subjectArtificial Intelligenceen
dc.subjectDigital Marketingen
dc.subjectClassificationen
dc.subjectPersonalized Customer Experienceen
dc.subjectPredictive Modelingen
dc.subjectMachine Learningen
dc.titleAnalyzing the impact of Artificial Intelligence in Digital Marketing on personalized customer experienceen
dc.typeThesisen
dc.typeHochschulschriftde
dc.identifier.doi10.34726/hss.2021.78000-
dc.publisher.placeWien-
tuw.thesisinformationTechnische Universität Wien-
tuw.publication.orgunitE194 - Institut für Information Systems Engineering-
dc.type.qualificationlevelDiploma-
dc.identifier.libraryidAC16189535-
dc.description.numberOfPages65-
dc.thesistypeDiplomarbeitde
dc.thesistypeDiploma Thesisen
item.openairecristypehttp://purl.org/coar/resource_type/c_18cf-
item.openairecristypehttp://purl.org/coar/resource_type/c_18cf-
item.languageiso639-1en-
item.grantfulltextopen-
item.fulltextwith Fulltext-
item.openairetypeThesis-
item.openairetypeHochschulschrift-
item.cerifentitytypePublications-
item.cerifentitytypePublications-
Appears in Collections:Thesis

Files in this item:

Show simple item record

Page view(s)

91
checked on Jun 14, 2021

Download(s)

79
checked on Jun 14, 2021

Google ScholarTM

Check


Items in reposiTUm are protected by copyright, with all rights reserved, unless otherwise indicated.