2 results about "Information gain ratio" patented technology

This invention belongs to the field of mechanical equipment condition monitoring and fault prediction technology, specifically relating to a method for constructing a health index for rolling bearings based on deep reinforcement learning. The method includes: acquiring features to be fused from sensor data of the mechanical equipment and performing normalization processing; utilizing deep reinforcement learning technology to simultaneously perform two stages of tasks: feature selection and feature weight allocation, to construct a comprehensive health index; wherein the action space of deep reinforcement learning includes discrete feature selection actions and continuous feature weight allocation actions, and the reward function is based on the information gain ratio of the health index; by training a deep reinforcement learning agent, it learns to select the most effective subset of features and allocate optimal weights, thereby generating a health index. This invention can automatically and adaptively construct health indices, avoiding the problems of relying on expert knowledge and manual design in traditional methods, and improving the quality and generalization ability of health indices.

This invention provides a method for assessing the situation in 1v1 close-range UAV games based on a multi-criteria random forest. First, it collects state parameters such as position, attitude, and velocity of both friendly and enemy UAVs and labels the situation categories. Then, based on these state parameters, it constructs four types of advantage features: distance advantage, azimuth advantage, velocity advantage, and altitude advantage, and obtains advantage feature vectors through normalization. Next, it uses a bagging ensemble strategy to generate multiple training subsets, randomly assigning one or more splitting criteria from information entropy, information gain ratio, Gini impurity, and chi-square test to each decision tree. Feature selection and node splitting are then performed according to the corresponding criteria, forming a multi-criteria random forest model composed of multiple splitting criteria. Finally, the real-time collected game state data is converted into advantage feature vectors and input into the multi-criteria random forest model. The game situation category and corresponding probability distribution are obtained through the voting results of multiple decision trees, achieving real-time intelligent assessment of the 1v1 close-range UAV game situation. This invention significantly improves situation recognition accuracy, adaptability to unbalanced situation distributions, and reliability in identifying dangerous situations, while maintaining low computational complexity.