Access restricted to higher education institution's students and staff
master's thesis
Segmentation of customers due to the use of the term deposit service in the bank by using data mining methods

Ivan Palinkaš (2016)
Metadata
TitleSegmentacija klijenata s obzirom na korištenje usluge oročenja sredstava u banci s pomoću metoda rudarenja podataka
AuthorIvan Palinkaš
Mentor(s)Marijana Zekić Sušac (thesis advisor)
Abstract
Ovaj rad se bavi upotrebom neuronskih mreža u analizi baze podataka banke kako bi se prepoznali budući korisnici jedne bankarske usluge, dugoročne oročene štednje. Podaci su prikupljeni s pomoću tele marketing kampanje koju je provela banka u razdoblju od 2008. do 2010. godine. Na kraju marketing kampanje, uspješnost pristanka na uslugu oročenja je bila 8% što je za banku nedovoljno učinkovito. U ovom radu će se prikazati kako se pomoću neuronskih mreža može povećati uspješnost marketing kampanja. Neuronske mreže pripadaju u metode rudarenja podataka i koriste se za problem klasifikacije i predviđanja. U sklopu ovog rada, koristeći neuronske mreže kao sredstvo traženja veza među podacima u bazi izgradit će se model za klasifikaciju klijenata u dvije skupine: (1) potencijalni korisnici usluge oročenja sredstava i (2) oni koji će odbiti uslugu oročenja sredstava u toj banci. Takav model može se koristiti za segmentaciju klijenata, a njegova uporaba može dovesti do veće stope odgovara na marketinške kampanje. Testirano je više modela neuronskih mreža promjenom raspodjele uzoraka, aktivacijskih funkcija i funkcija greške, te je na temelju stope točnosti klasifikacije izabran najuspješniji model koji se preporuča za uporabu. Uz prikaz rezultata svih istraživanja, prikazat će se i objasniti matrica konfuzije najbolje mreže i analiza osjetljivosti. Na kraju su dane smjernice za uporabu modela i za daljnja istraživanja.
Keywordsneural networks classification segmentation of clients telemarketing term deposits
Parallel title (English)Segmentation of customers due to the use of the term deposit service in the bank by using data mining methods
Committee MembersMarijana Zekić Sušac (committee chairperson)
Josip Mesarić (committee member)
Đula Borozan (committee member)
GranterSveučilište Josipa Jurja Strossmayera u Osijeku
Ekonomski fakultet u Osijeku
Lower level organizational unitsKatedra za kvanititavne metode i upravljanje
PlaceOsijek
StateCroatia
Scientific field, discipline, subdisciplineSOCIAL SCIENCES
Economics
Business Informatics
Study programme typeuniversity
Study levelgraduate
Study programmeBusiness economy; specializations in: Business Informatics
Study specializationBusiness Informatics
Academic title abbreviationmag.oec.
Genremaster's thesis
Language Croatian
Defense date2016-09-21
Parallel abstract (English)
This paper deals with the usage of neural networks in the database analysis in order to identify future banking services users with, long-term deposits. Campaigns have been lead and information has been collected with help of telemarketing in the period from 2008. to 2010. At the end of last marketing campaign, the successful rate of the long-term deposit consent was 8%, which is insufficiently effective for the bank. This paper will show how with neural networks, success of marketing campaigns can be increased. Neural networks are one of methods for data mining and they are used to solve problems of classification and prediction. As part of this work, using neural networks as a means of finding connections between data in the database, model will be built for classification of clients into two groups: (1) potential users of term deposits and (2) those who will refuse term deposits service in the bank. This model can be used for segmentation of clients, and its use can lead to higher rates of responding to a marketing campaign. Models of neural networks where tested by changing sample distribution, activation and error function, and based on the classification accuracy rate, the best neural network is determined for usage. With the presentation of the research results, the best neural network confusion matrix and sensitivity analysis will also be shown and explained. At the end, guidelines are given for the use of the model and for further research.
Parallel keywords (Croatian)neuronske mreže klasifikacija segmentacija klijenata tele marketing oročena štednja
Resource typetext
Access conditionAccess restricted to higher education institution's students and staff
Terms of usehttp://rightsstatements.org/vocab/InC/1.0/
URN:NBNhttps://urn.nsk.hr/urn:nbn:hr:145:036513
CommitterGordana Kradijan