-
Automatisierte Identifikation von Anlagethemen mit Momentum und News-Sentiment
Absolut Report
-
Abstract
Künstliche Intelligenz revolutioniert Investmentstrategien! Die Deka Investment nutzt KI, um Anlagethemen mit Momentum- und News-Sentiment-Analysen zu identifizieren. Das Ergebnis: Effiziente Strategien, die Markttrends frühzeitig erkennen und klare Outperformance erzielen. Einblicke von Dr. Stefan Salbrechter, Prof. Dr. Dominik Wolff und Dr. Ulrich Neugebauer im Absolut Report Spezial 2025. Mehr dazu im aktuellen Bericht.
More information
-
DS@GT at CheckThat! 2025: Evaluating Context and Tokenization Strategies for Numerical Fact Verification
-
Abstract
Numerical claims — statements involving quantities, comparisons, and temporal references — pose unique
challenges for automated fact-checking systems. In this study, we evaluate modeling strategies for veracity
prediction of such claims using the QuanTemp dataset and building our own evidence retrieval pipeline. We
investigate three key factors: (1) the impact of more evidences with longer input context windows using
ModernBERT, (2) the effect of right-to-left (R2L) tokenization, and (3) their combined influence on classification
performance. Contrary to prior findings in arithmetic reasoning tasks, R2L tokenization does not boost natural
language inference (NLI) of numerical tasks. A longer context window does also not enhance veracity performance
either, highlighting evidence quality as the dominant bottleneck. Our best-performing system achieves competitive
macro-average F1 score of 0.57 and places us among the Top-4 submissions in Task 3 of CheckThat! 2025. Our
code is available at https://github.com/dsgt-arc/checkthat-2025-numerical.
-
Guided Topic Modeling with Word2Vec: A Technical Note
-
Abstract
We propose GTM (Guided Topic Modeling), an algorithm that enables the fast and flexible generation of comprehensive topic clusters from (a pair of) seed words. The unsupervised algorithm performs clustering in the word-embedding space while offering the possibility to adjust the characteristics of the topic clusters via several hyperparameters. Applications for this methodology are information retrieval, classification and the calculation of various topic indices from news feeds.
More information
-
Return Prediction Models and Portfolio Optimization: Evidence for Industry Portfolios
Journal of Risk and Financial Management
-
Abstract
An essential motive for investing in commodities is to enhance the performance of portfolios traditionally including only stocks and bonds. We analyze the in-sample and out-of-sample portfolio effects resulting from adding commodities to a stock-bond portfolio for commonly implemented asset-allocation strategies such as equally and strategically weighted portfolios, risk-parity, minimum-variance as well as reward-to-risk timing, mean-variance and Black-Litterman. We analyze different commodity groups such as agricultural and livestock com-modities that currently are critically discussed. The out-of-sample portfolio analysis indicates that the attainable benefits of commodities are much smaller than suggested by previous in-sample studies. Hence, in-sample analyses, such as spanning tests, might exaggerate the ad-vantages of commodities. Moreover, the portfolio gains greatly vary between different types of commodities and sub-periods. While aggregate commodity indices, industrial and precious metals as well as energy improve the performance of a stock-bond portfolio for most asset-allocation strategies, we hardly find positive portfolio effects for agriculture and livestock. Consequently, investments in food commodities are not essential for efficient asset allocation.
More information
-
Portfolio Optimization with Sector Return Prediction Models
Journal of Risk and Financial Management
-
Abstract
We analyze return predictability for U.S. sectors based on fundamental, macroeconomic, and technical indicators and analyze whether return predictions improve tactical asset allocation decisions. We study the out-of-sample predictive power of individual variables for forecasting sector returns and analyze multivariate predictive regression models, including OLS, regularized regressions, principal component regressions, the three-pass regression filter, and forecast combinations. Using an out-of-sample Black–Litterman portfolio optimization framework and employing predicted returns as investors’ ‘views’, we evaluate the benefits of sector return forecasts for investors. We find that portfolio optimization with sector return prediction models significantly outperforms portfolios using historical averages as well as passive benchmark portfolios.
More information
-
Martingale defects in the volatility surface and bubble conditions in the underlying
19.01.2024
-
Abstract
The martingale theory of bubbles enables testing for asset price bubbles by analyzing option prices. As recently shown by Piiroinen et al. (Asset price bubbles: an option-based indicator, 2018), the SABR model is a strict local martingale when its parameterization implies a positive correlation between stock and option prices. We operationalize this theoretical result and analyze stock price bubbles in 2576 stocks over 26 years. Martingale defect conditions are absorbed quickly by options markets, but identify high proportions in significant and permanent changes in distribution of price returns, option trading activity, short interest in the underlying, and institutional ownership. These results confirm many common assumptions about stock price bubbles. These bubbles are temporally clustered, and tend to occur in periods of positive market development. Martingale defects are rare in market corrections, which indicates that they are a result of overoptimistic speculation.
More information
-
When Machines Trade on Corporate Disclosures: Using Text Analytics for Investment Strategies
Decision Support Systems Volume 165 , February 2023, 113892
-
Abstract
In this study, we evaluate several trading strategies based on the textual content of corporate disclosures using machine learning. To obtain a conservative estimate of profitability, we require orders to be placed for the close price of the current trading day and stocks must exhibit high liquidity to ensure proper order execution. Our evaluation based on 354,992 form 8-K filings and 10,204 ad hoc announcements shows that the proposed trading strategies yield up to 7.81% and 9.34% out-of-sample annualized return. More importantly, we find that the prevalent approach in the literature of estimating the stock market reaction of a disclosure based on the closing price of the past trading and omitting liquidity filters substantially overestimates profitability. We also provide useful insights for practitioners by describing feature importance to shed light onto how the machine learning models arrive at decisions.
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3910451
More information
-
Firm-specific Climate Risk Estimated from Public News
-
Abstract
We estimate firm-specific exposures to climate risk from public news covering a period of 20 years by applying a novel topic modeling algorithm. We differentiate between regulatory (or transition) and physical climate risks and document that financial markets price both risks. Our study is the first to find a positive and statistically significant risk premium for physical climate risk. For regulatory climate risk we find a regime shift occurring around the year 2012 reconciling the conflicting evidence in the literature. While the risk premium is positive in the earlier period, it becomes significantly negative in the later one. A long-short portfolio that is long “green” firms and short “brown” firms, as identified by their topic exposures in public news, constitutes a priced risk factor and shows a surprisingly strong correlation with an ESG-sorted benchmark portfolio.
More information
-
Vol, Skew and Smile Trading
The Journal of Derivatives
-
Abstract
Al-Jaaf, A., & Carr, P. (2023). Vol, Skew, and Smile Trading. The Journal of Derivatives 31:64-95.
We show how a three-strike option portfolio can be used to trade the difference between the instantaneous variance rate and the implied variance rate, the difference between the instantaneous covariation rate and the implied slope, or the difference between the instantaneous variance rate of volatility and the implied convexity. We label each one of these strategies as vol, skew, and smile trades. Our results yield precise financial interpretations of particular measures of the level, slope, and curvature of a BMS implied variance curve. We provide empirical evidence that the average returns of the vol and smile (skew) trades are negative (positive) and that the returns of the skew and smile trades cannot be explained by the CAPM.
© [2023] PMR. All rights reserved.
More information
-
Stock Picking with Machine Learning
Journal of Forecasting
-
Abstract
We analyze machine learning algorithms for stock selection. Our study builds on weekly data for the historical constituents of the S&P500 over the period from January 1999 to March 2021 and builds on typical equity factors, additional firm fundamentals, and technical indicators. A variety of machine learning models are trained on the binary classification task to predict whether a specific stock outperforms or underperforms the cross-sectional median return over the subsequent week. We analyze weekly trading strategies that invest in stocks with the highest predicted outperformance probability. Our empirical results show substantial and significant outperformance of machine learning-based stock selection models compared to an equally weighted benchmark. Interestingly, we find more simplistic regularized logistic regression models to perform similarly well compared to more complex machine learning models. The results are robust when applied to the STOXX Europe 600 as alternative asset universe.
More information
