Qiaoqi Lang, Jiqian Wang, Feng Ma, Dengshi Huang and Mohamed Wahab Mohamed Ismail
This paper verifies whether popular Internet information from Internet forum and search engine exhibit useful content for forecasting the volatility in Chinese stock market.
Abstract
Purpose
This paper verifies whether popular Internet information from Internet forum and search engine exhibit useful content for forecasting the volatility in Chinese stock market.
Design/methodology/approach
First, the authors’ study commences with several HAR-RV-type models, then the study amplifies them respectively with the posting volume and search frequency to construct HAR-IF-type and HAR-BD-type models. Second, from in-sample and out-of-sample analysis, the authors empirically investigate the interpretive ability, forecasting performance (statistic and economic). Third, various robustness checks are utilized to reconfirm the authors’ findings, including alternative forecast window, alternative evaluation method and alternative stock market. Finally, the authors further discuss the forecasting performance in different forecast horizons (h = 5, 10 and 20) and asymmetric effect of information from Internet forum.
Findings
From in-sample perspective, the authors discover that posting volume exhibits better analytical ability for Chinese stock volatility than search frequency. Out-of-sample results indicate that forecasting models with posting volume could achieve a superior forecasting performance and increased economic value than competing models.
Practical implications
These findings can help investors and decision-makers obtain higher forecasting accuracy and economic gains.
Originality/value
This study enriches the existing research findings about the volatility forecasting of stock market from two dimensions. First, the authors thoroughly investigate whether the Internet information could enhance the efficiency and accuracy of the volatility forecasting concerning with the Chinese stock market. Second, the authors find a novel evidence that the information from Internet forum is more superior to search frequency in volatility forecasting of stock market. Third, they find that this study not only compares the predictability of the posting volume and search frequency simply, but it also divides the posting volume into “good” and “bad” segments to clarify its asymmetric effect respectively.
Highlights
This study aims to verify whether posting volume and search frequency contain predictive content for estimating the volatility in Chinese stock market.
The forecasting model with posting volume can achieve a superior forecasting performance and increases economic value than competing models.
The results are robust in alternative forecast window, alternative evaluation method and alternative market index.
The posting volume still can help to forecast future volatility for mid- and long-term forecast horizons. Additionally, the role of posting volume in forecasting Chinese stock volatility is asymmetric.
Details
Keywords
Sukhmani Bhatia Chugh and Archana Goel
With the increase in uncertainty around the globe, an intensifying interest is seen in Economic Policy Uncertainty (EPU) as a topic of research. Researchers worldwide understand…
Abstract
With the increase in uncertainty around the globe, an intensifying interest is seen in Economic Policy Uncertainty (EPU) as a topic of research. Researchers worldwide understand the significance of the impact of EPU on the country's development. EPU has a far-reaching impact as uncertainty shocks in one part of the world resonate worldwide due to the level of interconnectivity, globalization and quick communication. In order to facilitate these researchers, this study presents a bibliometric analysis of the existing research in this field using VOS viewer software, by consolidating all the studies from Scopus indexed journal articles, conference proceedings and review papers published in English language from 2006 to 2022. Bibliometric analysis on EPU has rarely been performed. The analysis identifies the publication trends, journal-wise citation, most influential authors, countries, institutions, keyword co-occurrence and authors of different countries who have collaborated for the research in the field. Finally, 1,055 papers were used for bibliometric analysis. The findings depicted that the most cited article on EPU is ‘Measuring economic policy uncertainty’ by Baker et al. (2016) and the most prolific author appears to be Rangan Gupta from University of Pretoria which as an institution also has the maximum publications on this topic. The Journal Finance Research Letters has published the greatest number of researches on EPU. This chapter also summarizes the limitations of the study along with new areas of research.
Details
Keywords
Jiqian Dong, Sikai Chen, Mohammad Miralinaghi, Tiantian Chen and Samuel Labi
Perception has been identified as the main cause underlying most autonomous vehicle related accidents. As the key technology in perception, deep learning (DL) based computer…
Abstract
Purpose
Perception has been identified as the main cause underlying most autonomous vehicle related accidents. As the key technology in perception, deep learning (DL) based computer vision models are generally considered to be black boxes due to poor interpretability. These have exacerbated user distrust and further forestalled their widespread deployment in practical usage. This paper aims to develop explainable DL models for autonomous driving by jointly predicting potential driving actions with corresponding explanations. The explainable DL models can not only boost user trust in autonomy but also serve as a diagnostic approach to identify any model deficiencies or limitations during the system development phase.
Design/methodology/approach
This paper proposes an explainable end-to-end autonomous driving system based on “Transformer,” a state-of-the-art self-attention (SA) based model. The model maps visual features from images collected by onboard cameras to guide potential driving actions with corresponding explanations, and aims to achieve soft attention over the image’s global features.
Findings
The results demonstrate the efficacy of the proposed model as it exhibits superior performance (in terms of correct prediction of actions and explanations) compared to the benchmark model by a significant margin with much lower computational cost on a public data set (BDD-OIA). From the ablation studies, the proposed SA module also outperforms other attention mechanisms in feature fusion and can generate meaningful representations for downstream prediction.
Originality/value
In the contexts of situational awareness and driver assistance, the proposed model can perform as a driving alarm system for both human-driven vehicles and autonomous vehicles because it is capable of quickly understanding/characterizing the environment and identifying any infeasible driving actions. In addition, the extra explanation head of the proposed model provides an extra channel for sanity checks to guarantee that the model learns the ideal causal relationships. This provision is critical in the development of autonomous systems.