Random Forest and LightGBM-Based Human Health Check for Medical Device Fault Detection

Medical devices are items used directly or indirectly in the human body and are a prerequisite for hospital treatment of patients, and their quality can have a direct impact on the health of patients, so strengthening the quality control of medical device use is a hot spot of concern in the clinic. Current medical device testing can reduce the occurrence of adverse events, but it cannot be completely avoided, and its work still needs to be further strengthened. In this paper, we design a two-way feature selection algorithm based on PSO_RF. We use random forest to calculate the importance of the feature attributes of the sample data and sort the results in descending order, where a particle swarm algorithm is introduced to optimize the parameters of the random forest algorithm. The 245 medical device adverse event reports received by the testing center were selected, the occurrence and types of adverse events were analyzed retrospectively, and quality control countermeasures for medical device use were formulated.