Online examination integrated intelligent paper marking method and system
By constructing a subject knowledge graph and graph attention network, dynamically adjusting weights, and combining the BiLSTM-CRF model and dependency syntax analysis, the problem of insufficient logical chain parsing in online exam grading is solved, and accurate scoring of answers to subjective questions is achieved.
Patent Information
- Application Number
- CN202510915394.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-03
- Publication Date
- 2025-10-03
- Estimated Expiration
- 2045-07-03
AI Technical Summary
The existing online examination marking system lacks in-depth analysis of the integrity of the logical chain when judging subjective questions, is unable to handle the semantic disambiguation of subject professional terms, and the static weight allocation strategy cannot dynamically adjust the importance of test points, resulting in inaccurate scoring.
Construct a subject knowledge graph, dynamically assign entity weights, generate a logical coherence score through multi-granularity parsing and graph attention network, calculate the final score by combining entity coverage and logical coherence, use the BiLSTM-CRF model for semantic disambiguation and dependency syntax analysis, prune invalid associations, and construct a directed logical graph.
It achieves accurate analysis of the logical relationship of answers to subjective questions, improves the scientific nature and adaptability of scoring, can automatically identify logical relationships such as cause and effect, and transitions, dynamically adjust the importance of knowledge points, and improve the accuracy of scoring.