master's thesis
CLASSIFICATION OF EMPLOYEES WITH THE DECISION TREE AND NEURAL NETWORKS

Maja Lukačević (2017)
Sveučilište Josipa Jurja Strossmayera u Osijeku
Ekonomski fakultet u Osijeku
Katedra za kvanititavne metode i upravljanje
Metadata
TitleKlasifikacija zaposlenika s pomoću stabla odlučivanja i neuronskih mreža.
AuthorMaja Lukačević
Mentor(s)Marijana Zekić Sušac (thesis advisor)
Abstract
Predmet istraživanja u ovom radu je odlazak zaposlenika iz poduzeća kroz gradnju modela neuronskih mreža i stabla odlučivanja napravljenom u softverskom paketu Statistica. Ovaj rad pokušava razumjeti zašto zaposlenici napuštaju poduzeće, koji je glavni razlog, zašto zaposlenici donose takvu odluku i kako da se spriječi odljev talentiranih zaposlenika. Zaposlenici su podijeljeni u dvije grupe. Klasa 0 predstavlja zaposlenike u poduzeću, a klasa 1 predstavlja zaposlenika koji je napustio poduzeće. Modeli neuronske mreže koristili su dvije funkcije, logističku aktivacijsku funkciju i tangens aktivacijsku funkciju. Rezultati su pokazali da je logistička (94,07%) imala bolju stopu točnosti nego tangens (93,21%) aktivacijska funkcija. Stablo odlučivanja ima najbolju stopu 95,47%, zbog bolje stope taj model pokazao se kao uspješniji. Najvažnija varijabla zaposlenikova napuštanja u stablu odlučivanja je stupanj zadovoljstva, a najmanju važnost ima promaknuće u zadnjih 5 godina.
Keywordsdata mining decision tree neural networks employees turnover business intelligence
Parallel title (English)CLASSIFICATION OF EMPLOYEES WITH THE DECISION TREE AND NEURAL NETWORKS
Committee MembersMarijana Zekić Sušac (committee chairperson)
Davor Dujak (committee member)
Nataša Šarlija (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 date2017-04-12
Parallel abstract (English)
The subject of research in this study is modeling employees turnover by using neural networks and decision tree methods in software package Statistica. This study is trying to understand why employees leave, i.e. what are the main factors why people make that decision and how to prevent the drain of employee talent. Employees are classified in two classes. Class 0 represent employees who stayed in company and class 1 represent employees who have already left the company. Models of neural networks in this study use two functions: logistic activation function and tanh activation function. Results showed that the logistic (94,07%) have better rate then tanh (93,21%) activation function. Decision tree model produces the highest rate of 95,47% and is the the most successful model. The most important variable on employee turnover in decision tree model was the satisfaction level, while the lowest importance has promotion in last 5 years.
Parallel keywords (Croatian)rudarenje podataka stablo odlučivanja neuronske mreže odlazak zaposlenika poslovna inteligencija
Resource typetext
Access conditionAccess restricted to students and staff of home institution
Terms of usehttp://rightsstatements.org/vocab/InC/1.0/
URN:NBNhttps://urn.nsk.hr/urn:nbn:hr:145:617707
CommitterGordana Kradijan