Scholarlink

<p>The work introduces a hybrid ensemble model that combines conventional stock market prediction models with sentiment analysis of news articles in order to improve the accuracy of predictions. By utilizing the advantages of Long Short-Term Memory(LSTM), Gated Recurrent Unit(GRU), Bidirectional LSTM (BiLSTM), and Recurrent Neural Network(RNN) models in an ensemble framework, we attain an impressive average prediction accuracy of 91.89% where our model was evaluated on ten stocks and surpassed the performance of current models. This outcome underscores the need to integrate news sentiment with technical indicators to get a thorough comprehension of market dynamics. Moreover, the proposed model-driven portfolio regularly outperforms the Nifty 50 benchmark at different risk tolerance levels (0.3, 0.5, and 0.7), generating a stable positive alpha. This indicates greater returns when adjusted for risk. The model’s ability to adapt to the varying needs of investors is demonstrated by the performance it achieved across risk profiles. The proposed model is also compared with the existing models to show the model’s efficiency.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Ensemble time series models for stock price prediction and portfolio optimization with sentiment analysis

  • 期刊论文

  • 发表日期：

    2025年3月6日

查看全文
  • Malineni Lakshmi Narayana,
  • Arundhati J Kartha,
  • Ankur Kumar Mandal,
  • Roshini P,
  • Akshaya Suresh,
  • Arun Cyril Jose

摘要

The work introduces a hybrid ensemble model that combines conventional stock market prediction models with sentiment analysis of news articles in order to improve the accuracy of predictions. By utilizing the advantages of Long Short-Term Memory(LSTM), Gated Recurrent Unit(GRU), Bidirectional LSTM (BiLSTM), and Recurrent Neural Network(RNN) models in an ensemble framework, we attain an impressive average prediction accuracy of 91.89% where our model was evaluated on ten stocks and surpassed the performance of current models. This outcome underscores the need to integrate news sentiment with technical indicators to get a thorough comprehension of market dynamics. Moreover, the proposed model-driven portfolio regularly outperforms the Nifty 50 benchmark at different risk tolerance levels (0.3, 0.5, and 0.7), generating a stable positive alpha. This indicates greater returns when adjusted for risk. The model’s ability to adapt to the varying needs of investors is demonstrated by the performance it achieved across risk profiles. The proposed model is also compared with the existing models to show the model’s efficiency.