Signal sparse decomposition-based intelligent positioning method for single-phase earth fault section of power distribution network

The invention discloses a signal sparse decomposition-based intelligent positioning method for a single-phase earth fault section of a power distribution network, which comprises the following steps of: iteratively comparing weighted information gains by using a weighted ID3 algorithm to select an optimal characteristic to construct a weighted decision tree model, and inputting a training data set for training to obtain a trained decision tree model, the training data set is a zero-sequence current waveform transient signal database when the single-phase earth fault occurs after marking the best matching atoms, and the best matching atom acquisition mode is to realize atom matching pursuit by using a particle swarm algorithm; and judging whether a single-phase earth fault exists in the power distribution network system, recording a system zero-sequence current waveform transient signal through a field oscillograph, inputting the system zero-sequence current waveform transient signal into the trained decision