Metode dubokog učenja za analizu i predviđanje tržišta dionica

master's thesis

• Title (english): Deep learning methods for stock market analysis and prediction
• Author: Tomislav Medić
• Mentor: Mirjana Pejić Bach (mentor)
• Committee member: Ivan Strugar (predsjednik povjerenstva)
• Committee member: Jovana Zoroja (član povjerenstva)
• Granter: University of Zagreb; Faculty of Economics and Business; (Department of Informatics); Zagreb
• Defense date and country: 2019-09-25, Croatia
• Scientific / art field, discipline and subdiscipline: SOCIAL SCIENCES; Economics; Business Informatics

Abstract

U posljednjih nekoliko godina, svjedočimo nevjerojatnom razvoju umjetne inteligencije i njenih podvrsta poput strojnog i dubokog učenja. Po prvi puta u povijesti, pristup podacima i razvoj tehnologije omogućuju efikasnu primjenu umjetne inteligencije u različitim područjima, od samovoznih (engl. self driving) automobila do virtualnih asistenata, robota u skladištima (npr. Alibaba), robota na proizvodnim trakama, i sl.

Razvoj interneta potaknuo je razvoj različitih tehnologija koje su umjetnoj inteligenciji omogućile „skok“ sa teoretskih modela na praktične, uključujući: rudarenje podataka, razvoj programskih jezika (npr. Python) i knjižnica programskog koda, različitih alata poslovne inteligencije i sl. Danas vrlo jednostavno i brzo možemo napraviti umjetne neuronske mreže (duboko učenje) te ih primijeniti na različite analize i zadatke, a time metode dubokog učenja iz dana u dan pronalaze nova područja u kojima optimiziraju procese, smanjuju troškove i potiču inovacije.

Jedno od tih područja je i tržište vrijednosnih papira, gdje se duboko učenje koristi za različite analize, uočavanje trendova i predviđanje, a razvoj neuronskih mreža na tom području i nedavni uspjesi te tehnologije, predstavljaju poticaj za daljnje istraživanje i sve dublju primjenu metoda dubokog učenja na tržištu vrijednosnih papira.

Abstract (english)

In recent years, we are witnessing the incredible development of artificial intelligence and its sub-types such as machine learning and deep learning. For the first time in history, access to data and technological development enable the effective application of artificial intelligence in a variety of fields, from self-driving cars to virtual assistants, warehouse robots (eg. Alibaba), production lane robots, and the like.

The development of the Internet has encouraged the development of various technologies that have allowed artificial intelligence to „jump“ from theoretical models to practical ones, including: data mining, development of programming languages (eg. Python) and code libraries, various business intelligence tools, etc. Today, we can build artificial neural networks (deep learning) easily and quickly, and apply them to a variety of analyses and tasks, which have allowed deep learning to constantly find new areas where they optimize processes, reduce costs and drive innovation.

One of these areas is the securities market, where deep learning is used for various analyses, trend recognition and forecasting, and the development of neural networks in this area along with recent technological successes are an incentive for further research and deeper application of deep learning methods to securities market.

• Keywords: umjetna inteligencija; duboko učenje; umjetne neuronske mreže; tržište dionica; predviđanje
• Keywords (english): artificial intelligence; deep learning; artificial neural networks; stock market; forecasting; prediction
• Language: croatian
• URN:NBN: urn:nbn:hr:148:934056
• Study programme: Title: Business Economics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra ekonomije (magistar/magistra ekonomije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2019-09-27 12:38:08