Text‐based soybean futures price forecasting: A two‐stage deep learning approach

Abstract

This paper investigates the soybean futures price prediction problem from a new perspective and proposes an effective prediction model named Two‐Stage Hybrid Long Short‐Term Memory (TSH‐LSTM) by using text data from social media. First, the unstructured text is transformed into structured data by sentiment analysis and text classification methods. The improved sentiment score is computed by combining the degree centrality of sentiment words based on the sentiment dictionary method, and the characteristics of price fluctuations in texts are learned through the text Recurrent Convolutional Neural Networks. Second, the significant relationship between social media features and soybean futures price is assessed through stepwise regression, and the results of such an assessment are used as a basis for the identification of significant factors as input variables of the prediction model. Finally, the TSH‐LSTM prediction model is designed, and the final prediction result is acquired through the combination of prediction results of each stage using the error reciprocal method. The empirical results indicate that the incorporation of the social media text feature helps improve forecasting performances. Specifically, the proposed TSH‐LSTM is more accurate than univariate LSTM, multivariate LSTM, and eXtreme Gradient Boosting.

A deep learning model for online doctor rating prediction

“…Then for a word in where and , the concatenation of both forward and backward embeddings is performed in the form where and . For details on LSTM implementation, please refer to An et al (2022) and Kao et al (2021).…”

“…Other advantages include no requirement for explicit feature engineering (Weng et al, 2020) and data representation through non-linear transformations (Kraus & Feuerriegel, 2017). DL models have been applied to various NLP tasks, such as predicting soybean price (An et al, 2022), and bank question and answer (Wei & Liang, 2022). Despite the considerable advantages of DL in dealing with textual data, literature is scarce on the application of DL in online doctor rating prediction task.…”

Equivalent sentiment measures for cross-language analysis of corporate communications

“…What our method contributes to the state-of-the-art is a bridge between non-English texts and machine learning based on a transparent process. As mentioned above, English texts have been used in studies of global markets with machine-learning methods [ 19 ], where machine-translation is the most common approach to the study of non-English texts [ 20 ]. Text-based features sourced from machine-translation to English do increase forecast accuracy in practice [ 21 ].…”

Comprehensive commodity price forecasting framework using text mining methods

“…MLP networks are the most popular and frequently used models of ANNs in many practical applications (Massaoudi et al, 2021). LSTM is a deep learning model and it has been widely applied in time series forecasting (An et al, 2023; Jiao et al, 2022).…”