20.500.12556/RUL-164325 Razvoj in implementacija algoritmov za trgovanje s poudarkom na analizi časovnih vrst Development and implementation of trading algorithms focusing on time series analysis V svetu trgovanja se vedno išče nove načine, kako bi pridobili prednost pred drugimi trgovalci. Trgovalci zaradi tega uporabljajo različne metode analize časovnih vrst ter strategij. S pomočjo analize časovnih vrst lahko napovemo vrednosti časovne vrste, kar nam lahko pomaga pri odločanju v trgovalni strategiji. V okviru te magistrske naloge smo za analizo časovnih vrst uporabljali grafe vidljivosti, kar je povezava med analizo časovnih vrst in analizo grafov. Predstavili smo že obstoječe in nove pristope k napovedi časovnih vrst. Nekateri pristopi so bili uspešnejši od obstoječih, nekateri pa ne. S pomočjo grafov vidljivosti smo tudi predstavili nekaj indikatorjev in strategij, ki te indikatorje uporabljajo za trgovanje. Strategije so bile relativno zelo uspešne v primerjavi s strategijo ''kupi in drži'' ter tudi po drugih metrikah, kot je na primer razmerje Sharpe. Ena izmed strategij je v primerjavi s stratagijo ''kupi in drži'' zaslužila približno 4-krat več, pri čemer pa je ohranjala tudi manjšo volatilnost ter boljše razmerje Sharpe, ki je znašal 0,8. In the world of trading, there is always a search for new ways to gain an edge over other traders. As a result, traders employ various methods of time series analysis and strategies. By analyzing time series, we can forecast the value of a time series, which can assist in decision-making within a trading strategy. In this master's thesis, we used visibility graphs for time series analysis, which is a connection between time series analysis and graph analysis. We presented both existing and new approaches to time series forecasting. Some approaches were more successful than the existing ones, while others were less so. With the help of visibility graphs, we also introduced several indicators and strategies that use these indicators for trading. The strategies were highly successful relative to the buy-and-hold strategy, as well as according to other metrics, such as the Sharpe ratio. One of the strategies earned approximately four times more compared to the buy-and-hold strategy, while also maintaining lower volatility and a better Sharpe ratio, which was 0.8. analiza časovnih vrst algoritmično trgovanje trgovalne strategije graf vidljivosti time series analysis algorithmic trading trading strategies visibility graph true false false Slovenski jezik Angleški jezik Magistrsko delo/naloga 2024-10-22 13:05:02 2024-10-22 13:05:07 2024-11-25 11:20:28 0000-00-00 00:00:00 2024 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 37047 216315139 Pajer_Ivo_-_Razvoj_in_implementacija_algoritmov_za_trgovanje_s_poudarkom_na_analizi_casovnih_vrs.pdf Pajer_Ivo_-_Razvoj_in_implementacija_algoritmov_za_trgovanje_s_poudarkom_na_analizi_casovnih_vrs.pdf 1 C8A53C61B7BB2DEF471F9731B237BEFB 241a913e703ddaab3e5f06a84008d270fce4e8339629ce49ab294200540de4d2 353e94d6-9065-11ef-b232-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=194647 Fakulteta za računalništvo in informatiko 0 0 0