Procjena trenda u financijskim vremenskim nizovima zasnovana na strojnom učenju

undergraduate thesis

• Title (english): Machine Learning Based Trend Assessment in Financial Time Series
• Author: Petar Jukić
• Mentor: Zvonko Kostanjčar (mentor)
• Committee member: Zvonko Kostanjčar (predsjednik povjerenstva)
• Committee member: Stjepan Begušić (član povjerenstva)
• Committee member: Tomislav Kovačević (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-09, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Rad se bavi procjenom trenda u financijskim vremenskim nizovima zasnovano na strojnom učenju. Pregledom dosadašnjih znanstvenih radova na tu temu, prepoznati su problemi definicije uzlaznog/silaznog trenda, tj. oznaka, kao i problemi određivanja značajki i odabira adekvatnog modela. U radu je nad vremenskim nizom cijena dionica konstruiran posmični prozor koji za generiranje značajki i oznaka. U tako definiranom eksperimentalnom postavu, koristeći model nižeg kapaciteta, logističke regresije, isprobani su i uspoređeni različiti načini označavanja trenda. Koristeći odabrani indikator konvergencije/divergencije pomičnog prosjeka, treniran je stroj potpornih vektora, klasifikator većeg kapaciteta. Treniranje i validacija modela napravljena je na povijesnim cijenama dionica Njujorške burze. Uz zadovoljavajuće performanse modela strojnog učenja, predložena strategija trgovanja koja se oslanja na izlaz modela.

Abstract (english)

The paper deals with trend assessment in financial time series based on machine learning. A review of previous scientific work on this topic has revealed problems in defining an upward/downward trend (labels), as well as problems in determining features and an adequate model. A rolling window was constructed to generate features and labels. In the experimental setup thus defined, using a lower capacity model (logistic regression), different ways of labelling the trend were identified and compared. Using the selected moving average convergence/divergence indicator, a higher capacity support vector classifier was trained and validated. The training and validation of the model was performed using NYSE historical stock prices. Upon satisfactory performance of the machine learning model, a trading strategy was proposed based on the trained model output.

• Keywords: procjena trenda; financijski vremenski nizovi; stroj potpornih vektora; klasifikacija; strojno učenje
• Keywords (english): trend assessment; financial time series; support vector machine; classification; machine learning
• Language: croatian
• URN:NBN: urn:nbn:hr:168:137373
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-14 20:43:18