Paper: arXiv 2508.16378
Authors: Qizhao Chen
Abstract
This paper presents a dynamic cryptocurrency portfolio optimization strategy that integrates technical indicators and sentiment analysis to enhance investment decision-making. The proposed method employs the 14-day Relative Strength Index (RSI) and 14-day Simple Moving Average (SMA) to capture market momentum, while sentiment scores are extracted from news articles using the VADER (Valence Aware Dictionary and sEntiment Reasoner) model, with compound scores quantifying overall market tone. The large language model Google Gemini is used to further verify the sentiment scores predicted by VADER and give investment decisions. These technical indicator and sentiment signals are incorporated into the expected return estimates before applying mean-variance optimization with constraints on asset weights. The strategy is evaluated through a rolling-window backtest over cryptocurrency market data, with Bitcoin (BTC) and an equal-weighted portfolio of selected cryptocurrencies serving as benchmarks. Experimental results show that the proposed approach achieves a cumulative return of 38.72, substantially exceeding Bitcoin’s 8.85 and the equal-weighted portfolio’s 21.65 over the same period, and delivers a higher Sharpe ratio (1.1093 vs. 0.8853 and 1.0194, respectively). However, the strategy exhibits a larger maximum drawdown (-18.52%) compared to Bitcoin (-4.48%) and the equal-weighted portfolio (-11.02%), indicating higher short-term downside risk. These results highlight the potential of combining sentiment and technical signals to improve cryptocurrency portfolio performance, while also emphasizing the need to address risk exposure in volatile markets.
Complexity vs Empirical Score
- Math Complexity: 4.0/10
- Empirical Rigor: 7.5/10
- Quadrant: Street Traders — empirical and applied, lower math complexity
Why this score: The paper uses standard mean-variance optimization (MVO) and basic signal processing (RSI, SMA) without deep mathematical derivations, keeping complexity moderate. It scores high on empirical rigor due to a detailed five-year rolling-window backtest against benchmarks, reporting multiple financial metrics (cumulative return, Sharpe ratio, max drawdown).
Research Flowchart
flowchart TD
A["Research Goal: Enhance Cryptocurrency Portfolio<br>Optimization using Sentiment & Technical Signals"] --> B["Data Collection & Feature Extraction"]
B --> C["Computational Processing"]
C --> D["Portfolio Optimization"]
D --> E["Backtesting & Evaluation"]
E --> F["Key Findings & Outcomes"]
subgraph B ["Inputs"]
B1["Historical Price Data"]
B2["News Articles"]
end
subgraph C ["Signal Generation"]
C1["Technical Indicators<br>14-day RSI & SMA"]
C2["Sentiment Analysis<br>VADER Model"]
C3["LLM Verification<br>Google Gemini"]
end
subgraph D ["Model"]
D1["Mean-Variance Optimization<br>with Constraints"]
end
subgraph E ["Validation"]
E1["Rolling-Window Backtest"]
end
subgraph F ["Results"]
F1["High Returns (38.72%)"]
F2["High Sharpe Ratio (1.1093)"]
F3["Higher Max Drawdown<br>(-18.52%)"]
end
B1 & B2 --> C1
B2 --> C2
C2 --> C3
C1 & C3 --> D1
D1 --> E1
E1 --> F1
E1 --> F2
E1 --> F3
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a dynamic cryptocurrency portfolio strategy that integrates technical indicators and sentiment analysis to enhance investment decision-making.
Source sentence
This paper proposes a dynamic cryptocurrency portfolio strategy that integrates technical indicators and sentiment analysis to enhance investment decision-making.
Assumes 5
- Investors are rational and risk‑averse in the mean‑variance framework. model
Source sentence
The mean-variance framework assumes that investors are rational and risk‑averse, preferring portfolios that offer higher returns with lower risk.
- Portfolio must be fully invested and long‑only. execution
Source sentence
First, the portfolio must remain fully invested and long‑only
- Daily turnover is limited to 80% of portfolio notional. execution
Source sentence
= 0.80, limiting daily reallocation to at most 80% of the portfolio’s notional value.
- One‑way transaction cost is 0.1% per trade. execution
Source sentence
Let the one‑way transaction cost rate be c (0.1% in this study).
Show 1 more
- Risk model uses a 180‑day rolling covariance without shrinkage. model
Source sentence
the risk model itself relies on a 180‑day rolling covariance without shrinkage or robust adjustments
Data
cryptocurrency market · crypto · daily
- Cryptocurrency price and volume data market-prices public
Source sentence
All price and volume data are obtained using the Crypto Compare API1, which provides reliable and comprehensive market data across a wide range of digital assets.
- News article sentiment data text-news unknown
Source sentence
sentiment signals are extracted from news articles with VADER and further validated using the Google Gemini large language model.
Costs
Modeled
One‑way transaction cost rate of 0.1% applied to the ℓ1‑norm of weight changes each rebalancing.
Source sentence
Let the one‑way transaction cost rate be c (0.1% in this study). The total cost of rebalancing at time t is modeled as: TCost_t = c · \|w_t - w_{t-1}\|_1
Finds 3
Sharpe 1.11 (benchmark 0.89) · total return 38.7% · max drawdown 18.5%
- The sentiment‑enhanced strategy achieves higher cumulative return and higher Sharpe ratio than baselines such as Bitcoin buy‑and‑hold, equal‑weighted, and momentum portfolios. performance
Source sentence
The backtest results demonstrate that the proposed strategy consistently outperforms traditional benchmarks, including Bitcoin buy-and-hold, equal-weighted, and momentum‑based portfolios, in both cumulative returns and risk‑adjusted performance.
- Incorporating sentiment alongside technical indicators yields more consistent performance gains compared to using either alone. performance
Source sentence
comparing the sentiment‑only and technical‑only strategies shows that incorporating sentiment information alongside technical indicators can lead to more consistent performance gains.
- All strategies, including the proposed one, experience relatively large drawdowns during periods of market stress. negative
Source sentence
all strategies, including the proposed one, experience relatively large drawdowns, indicating that the signals used may not fully protect the portfolio during sharp market declines.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 19 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (22)
- VADER: A Parsimonious Rule-Based Model for Sentiment Analysis of Social Media Text (2014) — link
- Cryptocurrency-portfolios in a mean-variance framework (2019) — link
- Comparative Study of Portfolio Optimization Models for Cryptocurrency and Stock Markets (2025) — link
- Bitcoin Price Prediction Using Sentiment Analysis and Empirical Mode Decomposition (2025) — link
- Sentiment-Driven Bitcoin Price Range Forecasting: Enhancing CART Decision Trees With High-Dimensional Indicators and Twitter Dynamics (2025) — link
- Stock Price Prediction Using LLM-Based Sentiment Analysis (2024) — link
- Cryptocurrency Price Prediction using Forecasting and Sentiment Analysis (2022) — link
- BERT ’s sentiment score for portfolio optimization: a fine-tuned view in Black and Litterman model (2022) — link
and 14 more
- A Novel Approach to Portfolio Construction: An Application of FinBERT Sentiment Analysis and Credibilistic CVaR Criterion (2025) — link
- Deep Leaming for Stock Market Prediction Using Technical Indicators and Financial News Articles (2018) — link
- Stock Market Prediction System Using LSTM with Technical Indicators as Voters (2023) — link
- Stock Price Prediction with LLM-Guided Market Movement Signals and Transformer Model (2025) — link
- Stock Price Change Prediction Using Prompt-Based LLMs with RL-Enhanced Post-Hoc Adjustments (2025) — link
- Explore the Use of Prompt-Based LLM for Credit Risk Classification (2025) — link
- Sentiment-Aware Stock Price Prediction with Transformer and LLM-Generated Formulaic Alpha (2025) — link
- Optimal Markowitz portfolio using returns forecasted with time series and machine learning models (2025) — link
- Investment Portfolio Optimization Based on Modern Portfolio Theory and Deep Learning Models — → 2508.14999
- A novel sentiment correlation-based method with dual transformer model for stock price prediction (2025) — link
- Financial Sentiment Analysis: Techniques and Applications (2024) — link
- Qwen-Fin: Chinese Financial Sentiment Analysis Based on Fine-Tuned Large Language Model (2024) — link
- Out-of-Sample Equity Premium Prediction: Combination Forecasts and Links to the Real Economy (2009) — link
- Predicting Future Stock Prices Using Historical Data: A Machine Learning Approach (2025) — link
1 of 22 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.