The invention relates to the technical field of power grids, and discloses an
artificial intelligence prediction model construction method applying full-service
big data of a power distribution network, which is characterized by comprising the following steps of: 1) data preprocessing: selecting service data of power distribution network services for analysis and
verification, and dividing an overall
data set into a
training set and a
test set, wherein the characteristics are
transaction data and technical indexes, the
transaction data needs to be subjected to normalization
processing, the technical indexes are obtained through calculation of the
transaction data, and a normalization process is included in a calculation process. According to the
artificial intelligence prediction model construction method applying the power distribution network full-service
big data, power distribution network reliability weak points, high-loss points and potential safety
hazard points can be accurately positioned and graded, power distribution network fatigue aging and potential quality
hazard equipment are rapidly sorted, operation and maintenance key lines and equipment are comprehensively mastered, and a guidance basis is provided for a power distribution network technical improvement scheme.