Uporaba kazalnikov tehnične in fundamentalne analize za napovedovanje cen delnic s pomočjo strojnega učenja

Poredoš, Kevin (author), Hovelja, Tomaž (mentor)

V diplomski nalogi je raziskana uporaba umetne inteligence za napovedovanje gibanja cen na finančnih trgih. Glavni problem je, kako z uporabo zgodovinskih fundamentalnih in tehničnih kazalnikov izboljšati natančnost napovedi cen. Avtor je uporabil različne modele strojnega učenja, vključno z RNN, LSTM in CNN, ter analiziral njihove rezultate. Najboljši modeli so dosegli znatno višjo donosnost in boljše tveganje kot tradicionalna strategija Buy \& Hold. Naloga tako dokazuje potencial naprednih modelov pri optimizaciji investicijskih odločitev.

• Language: Slovenian
• Keywords: Strojno učenje, Tehnična analiza, Fundamentalna analiza, Napovedovanje cen delnic, LSTM, CNN, Hibridni model, Finančni kazalniki, Tehnični kazalniki, Časovne vrste
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2025
• PID: 20.500.12556/RUL-171861
• COBISS.SI-ID: 248529667
• Publication date in RUL: 03.09.2025

Secondary language

• Language: English
• Title: The use of technical and fundamental analysis indicators for stock price prediction using machine learning

Abstract

This thesis explores the application of artificial intelligence for predicting price movements in financial markets. The main challenge addressed is improving price prediction accuracy by utilizing historical fundamental and technical indicators. Various machine learning models, including RNN, LSTM, and CNN, were implemented and evaluated. The best-performing models demonstrated significantly higher returns and better risk management compared to the traditional Buy \& Hold strategy. This work highlights the potential of advanced models to optimize investment decisions.

• Keywords: Machine learning, Technical analysis, Fundamental analysis, Stock price prediction, LSTM, CNN, Hybrid model, Financial indicators, Technical indicators, Time series