Analysis of Subjective Question Correction Based on OM/QM Intelligent Marking System

The intelligent marking technology can effectively improve and solve the problems existing in online manual marking due to its high efficiency, high accuracy, and high reliability. However, there are still many problems in the technology, process, function, and safety of intelligent marking of subjective questions, which need to be improved urgently. This research deeply excavates and investigates the cases and personnel related to bright marking and puts forward corresponding suggestions based on specific problems while obtaining first-hand information on intelligent marking of subjective questions. Existing issues in brilliancy marking of personal questions: insufficient text recognition accuracy, complex recognition model training process; philosophical algorithm logic are complicated to “override”, and data are not traceable; the intelligent marking model is single, lacking a replication mechanism. Based on three problems, the following solutions are proposed: text recognition model based on detection-recognition correction mechanism; intelligent sample selection mechanism based on text clustering analysis; new smart marking mode based on the fusion of manual marking and machine marking.