Line fault cause identification feature quantity preference based on quantum genetic algorithm and support vector machine

Currently, intelligent analysis methods are widely used to identify the causes of transmission line faults. However, when the number of input feature dimensions is high, issues such as feature redundancy, interference from invalid features, and overfitting in model training can arise. This paper introduces a fault cause feature selection method that integrates quantum genetic algorithms with support vector machines. By constructing a feature selection model using quantum encoding, this method utilizes quantum rotation gates and probability amplitude adjustments to efficiently search for feature subsets, thereby retaining key identification information while reducing the input dimensionality. Experiments show that this method can reduce the required feature dimensions for identifying causes of faults like lightning strikes and wind-induced biases by 52.1%, and improve classification accuracy by 3.1%. This provides more accurate and effective input feature support for improving the accuracy of fault cause identification using intelligent methods.