BERT-based Financial Sentiment Index and LSTM-based Stock Return Predictability

Hiew, Joshua Zoen Git; Huang, Xin; Mou, Hao; Li, Duan; Wu, Qi; Xu, Yabo

doi:10.48550/arxiv.1906.09024

Cited by 16 publications

(20 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…BERT is a Transformer-based language model. An example of a study using BERT is Hiew et al ( 2019 ) where text data is given polarity by BERT using Weibo, a Chinese Social Networking Service, and stock price predictions are made using LSTM.…”

Section: Related Workmentioning

confidence: 99%

Constructing Equity Investment Strategies Using Analyst Reports and Regime Switching Models

Taguchi

Watanabe

Sakaji

et al. 2022

Front. Artif. Intell.

View full text Add to dashboard Cite

This study demonstrates whether analysts' sentiments toward individual stocks are useful for stock investment strategies. This is achieved by using natural language processing to create a polarity index from textual information in analyst reports. In this study, we performed time series forecasting for the created polarity index using deep learning, and clustered the forecasted values by volatility using a regime switching model. In addition, we constructed a portfolio from stock data and rebalanced it at each change point of the regime. Consequently, the investment strategy proposed in this study outperforms the benchmark portfolio in terms of returns. This suggests that the polarity index is useful for constructing stock investment strategies.

show abstract

Section: Related Workmentioning

confidence: 99%

Constructing Equity Investment Strategies Using Analyst Reports and Regime Switching Models

Taguchi

Watanabe

Sakaji

et al. 2022

Front. Artif. Intell.

View full text Add to dashboard Cite

show abstract

“…In the context of financial applications, it has been mainly applied for sentiment analysis: One of such applications is FinBERT [2], a variation of BERT [11] specialized to financial sentiment analysis; it has obtained state-of-the-art results on FiQA sentiment scoring and Financial PhraseBank benchmaks. In [25] provide a similar application but feeding the sentiment analysis index generated by BERT in a LSTM-based trading strategy to predict stock returns. • Few-shot Learning [16,22,52]: it is an extreme form of inductive learning, with very few examples (sometimes only one) being used to learn the target task model.…”

Section: Transfer Learning: Sub-paradigmsmentioning

confidence: 99%

“…All of these details are better outlined in the next section. News-rich to news-poor stocks [34]; mitigating class imbalance in credit scoring [33,49] [ 28,40] Feature Find a suitable feature mapping to approximate the source domain to the target domain Sentiment feature space [34]; Portfolio selection factors [27] [ 57,54] Parameter Learn shareable parameters or priors between the source and target tasks models BERT specialized to financial sentiment analysis [2,25]; Stock selection, forecasting [20,7]; yield curve forecasting [37] [11, 9] Relational-knowledge Learn a logical relationship or rules in the source domain and transfer it to the target domain [39,50] 3 Methodology…”

Section: Transfer Learning: Sub-paradigmsmentioning

confidence: 99%

QuantNet: Transferring Learning Across Systematic Trading Strategies

Koshiyama,

Flennerhag,

Blumberg

et al. 2020

Preprint

View full text Add to dashboard Cite

In this work we introduce QuantNet: an architecture that is capable of transferring knowledge over systematic trading strategies in several financial markets. By having a system that is able to leverage and share knowledge across them, our aim is two-fold: to circumvent the so-called Backtest Overfitting problem; and to generate higher risk-adjusted returns and fewer drawdowns. To do that, QuantNet exploits a form of modelling called Transfer Learning, where two layers are market-specific and another one is market-agnostic. This ensures that the transfer occurs across trading strategies, with the market-agnostic layer acting as a vehicle to share knowledge, cross-influence each strategy parameters, and ultimately the trading signal produced. In order to evaluate QuantNet, we compared its performance in relation to the option of not performing transfer learning, that is, using market-specific old-fashioned machine learning. In summary, our findings suggest that QuantNet performs better than non transfer-based trading strategies, improving Sharpe ratio in 15% and Calmar ratio in 41% across 3103 assets in 58 equity markets across the world. Code coming soon.

show abstract

“…In the SemEval 2017 SubTask 4 on sentiment analysis in Twitter, systems that utilized deep learning based architectures such as recurrent neural networks (RNN; Rumelhart et al, 1986) and long-short term memory (LSTM; Hochreiter & Schmidhuber, 1997;Graves & Schmidhuber, 2005) with pretrained word-embeddings such as GloVe (Pennington et al, 2014) were placed among the top performing systems (Cabanski et al, 2017;Ghosal et al, 2017;Mansar et al, 2017;Moore & Rayson, 2017). Further, recent research utilizing pretrained context-based representations of text such as the bidirectional encoder representations from transformers (BERT; Devlin et al, 2018) fine-tuned to the financial domain using Financial PhraseBank (Malo et al, 2014) outperform previously best performing models in sentiment analysis (Araci, 2019;Hiew et al, 2019). In this work, we utilize three models based on BERT-finBERT (Araci, 2019), DistilBERT (Sanh et al, 2019), and RoBERTa (Liu et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

SEntFiN 1.0: Entity‐aware sentiment analysis for financial news

Sinha

Kedas

Kumar

et al. 2022

Asso for Info Science & Tech

View full text Add to dashboard Cite

Fine-grained financial sentiment analysis on news headlines is a challenging task requiring human-annotated datasets to achieve high performance. Limited studies have tried to address the sentiment extraction task in a setting where multiple entities are present in a news headline. In an effort to further research in this area, we make publicly available SEntFiN 1.0, a humanannotated dataset of 10,753 news headlines with entity-sentiment annotations, of which 2,847 headlines contain multiple entities, often with conflicting sentiments. We augment our dataset with a database of over 1,000 financial entities and their various representations in news media amounting to over 5,000 phrases. We propose a framework that enables the extraction of entity-relevant sentiments using a feature-based approach rather than an expression-based approach. For sentiment extraction, we utilize 12 different learning schemes utilizing lexicon-based and pretrained sentence representations and five classification approaches. Our experiments indicate that lexicon-based N-gram ensembles are above par with pretrained word embedding schemes such as GloVe. Overall, RoBERTa and finBERT (domain-specific BERT) achieve the highest average accuracy of 94.29% and F1-score of 93.27%. Further, using over 210,000 entity-sentiment predictions, we validate the economic effect of sentiments on aggregate market movements over a long duration.

show abstract

BERT-based Financial Sentiment Index and LSTM-based Stock Return Predictability

Cited by 16 publications

References 0 publications

Constructing Equity Investment Strategies Using Analyst Reports and Regime Switching Models

Constructing Equity Investment Strategies Using Analyst Reports and Regime Switching Models

QuantNet: Transferring Learning Across Systematic Trading Strategies

SEntFiN 1.0: Entity‐aware sentiment analysis for financial news

Contact Info

Product

Resources

About