Early Warning System for Scientific Research Integrity based on Abnormal Data Recognition Algorithm
Wei Li · 2024
This study presents an innovative Early Warning System (EWS) aimed at maintaining the integrity of scientific research through the implementation of anomaly detection algorithms. The EWS is designed to proactively detect anomalies in both textual and image data, thereby facilitating timely intervention and guidance within scientific research management platforms. Through rigorous experimentation and analysis, the effectiveness of the proposed system was evaluated using a diverse dataset of $\mathbf{10,000}$ images and 20,000 text passages. The experimental results showed promising results, with text recognition accuracy ranging from $\mathbf{98.33 \%}$ to $\mathbf{98.72 \%}$ and image recognition accuracy ranging from 95.42% to 95.80% across 20 trials. These results indicate the robustness and reliability of the system in identifying similarities and anomalies in scientific research data. In addition, the experiments revealed a consistent trend of text recognition outperforming image recognition, underscoring the importance of text analysis in research integrity management.