Istraživanje ponašanja kretanja cijena dionica primjenom dubinske analize podataka

scientific master's theses

• Title (english): Investigation of stock price behaviour patterns using data mining
• Author: Željana Bošnjak Cihlar; VIAF: 305673757
• Mentor: Damir Kalpić (mentor)
• Committee member: Bojana Dalbelo-Bašić (predsjednik povjerenstva)
• Committee member: Damir Kalpić (član povjerenstva)
• Committee member: Robert Manger (član povjerenstva)
• Committee member: Mirta Baranović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Applied Mathematics); Zagreb
• Defense date and country: 2006-06-26, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing; Artificial Intelligence
• Universal decimal classification (UDC): 004; GENERALLY; Computer science and technology. Computing. Data processing

Abstract

Dubinska analiza podataka zadnjih se godina sve više razvija i primjenjuje u različitim područjima, a osobito je interesantna njezina primjena na tržištu vrijednosnih papira. Poznata svjetska tržišta vrijednosnih papira imaju dugu tradiciju pa kretanja cijena dionica prate utjecaje industrijskih sektora i različite druge trendove. U Hrvatskoj tržište vrijednosnih papira nema dugu tradiciju i tek se razvija. Zato se u ovom radu metode dubinske analize primijenilo na podatke s hrvatskog tržišta vrijednosnih papira kako bi se doznalo postoje li zakonitosti ili ponašanja slična onima kakva su utvrđena na tradicijskim tržištima. Rezultati primjene metoda analize skupina su pokazali da postoje grupe dionica koje se slično ponašaju, ali također i da se dobivene skupine ne poklapaju s djelatnostima izdavatelja, tj. da se takav utjecaj nije znakovito očitovao na skupu analiziranih dionica. Također se pokazalo da bolje modele daju skupovi podataka s pomičnim prosječnim cijenama i postotcima promjene cijena, nego skupovi sa samim cijenama. Uspoređena je primjenjivost različitih metoda i algoritama. Očitovane razlike daju prednost nekim metodama, ali kvalitativno dobiveni rezultati općenito upućuju na jednake zaključke. Analiza periodičnosti nije pokazala postojanje punih periodičnosti na bazi tjedna, ali je pokazala neka interesantna pravila.

Abstract (english)

Data mining has attracted attention in different areas in recent years; especially its application has developed in finance and stock markets. Price movements of stocks traded on world stock exchanges with long tradition follow specific and known trends and are mainly influenced by factors specific to the particular industry sector to which this stock belongs. Croatian stock market is young and still developing. In this work data from Croatian stock market was used and data mining methods applied to find patterns and behaviour similar to ones on international stock markets. Cluster analysis showed existence of different stock clusters but obtained clusters did not correspond with globally accepted industry sector classifications. Analysis using moving average prices or price change ratio data produced better models than raw price data. Application of different analysis methods and algorithms was analysed. Some methods produced better results than the others, but differences did not change overall conclusions. Association analysis did not show periodicity on a weekly basis, although some interesting rules were discovered.

• Keywords: dubinska analiza podataka; cijene dionica; klasifikacija; analiza skupina; WEKA; vremensko-slijedni podatci
• Keywords (english): data mining; stock prices; clustering; classification; WEKA; time series data
• Language: croatian
• URN:NBN: urn:nbn:hr:168:782059
• Study programme: Title: Postgraduate master programme in electrical engineering; Study programme type: university; Study level: postgraduate; Academic / professional title: Magistar znanosti elektrotehnike (Magistar znanosti elektrotehnike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-07-11 11:42:05