Please use this identifier to cite or link to this item:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/84204| Title: | 套用部位管理於被動攻擊型線上投資組合策略之研究 Applying Position Sizing to Passive-Aggressive Algorithms for Online Portfolio Selection |
| Authors: | Chen-Yu Wang 汪宸宇 |
| Advisor: | 呂育道(Yuh-Dauh Lyuu) |
| Keyword: | 線上投資組合選擇,投資組合最佳化,被動攻擊算法,部位管理,條件機率,量化交易, online portfolio selection,portfolio optimization,passive-aggressive algorithms,position sizing,conditional probability,quantitative trading, |
| Publication Year : | 2022 |
| Degree: | 碩士 |
| Abstract: | 投資組合最佳化和部位管理是兩個投資中重要但不同的概念,過去針對投資組合選擇的研究,大多為可以及時處理最新資料、且無法預見未來的線上算法,且大多不會考慮部位管理,使所有資產在任何情況下都百分之百投入標的物,這不一定是對長期報酬最佳的選項。本研究對既有的被動攻擊型投資組合策略進行改寫,得到具備部位管理功能的新算法架構。在此一架構底下,使用者可以自行設計部位管理策略、或投資組合最佳化目標,形成自己的投資組合最佳化策略。基於此架構,我們接著以資料學習的方式找尋適當且單純的條件,作為部位調整的根據,套用於四個既有的被動攻擊型投資組合策略之上。實證研究顯示,在測試的六個資料集中,使用部位管理後,這些策略都能在其中的五個資料集上產生更高的最終獲利,並降低最大回檔比例。 Portfolio optimization and position sizing are two distinct and crucial concepts in investing. Past studies on portfolio selection are often online algorithms, meaning they process only past information in order to make decisions; past studies mostly assumed that captial is fully invested, which might be harmful to long-term returns. This thesis adjusts the formulas of some passive-aggressive algorithms for online portfolio selection to make them capable of incorporating a position sizing scheme. Then, to achieve better performance, an investor can develop their own position sizing strategy or their own optimization criteria for portfolio allocation. We then propose a simple position sizing strategy, based on insights from inspecting six benchmark datasets, to obtain better trading results in terms of terminal wealth and maximum drawdown when applied to four existing passive-aggressive algorithms. |
| URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/84204 |
| DOI: | 10.6342/NTU202200720 |
| Fulltext Rights: | 同意授權(限校園內公開) |
| metadata.dc.date.embargo-lift: | 2022-07-05 |
| Appears in Collections: | 資訊網路與多媒體研究所 |
Files in This Item:
| File | Size | Format | |
|---|---|---|---|
| U0001-2504202216463100.pdf Access limited in NTU ip range | 711.66 kB | Adobe PDF |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
