The invention discloses a demand
prediction system of construction materials of a power distribution network. The
system comprises a
data acquisition module, a clustering module, a
project type attribute determining module and a material demand prediction module; the
data acquisition module obtains parameters of preset attributes of historical
power grid projects as well as the material utilization amount of the different types of historical
power grid projects; the clustering module carries out clustering analysis on the material utilization amount of the historical
power grid projects in an improved k-means clustering method, and determines a clustering family; the
project type attribute determining module uses a key word
frequency analysis method in the clustering family to determine a
project type attribute; and according to the material utilization amount of the historical power grid project, the parameter of the preset attribute of the historical project and the project
type attribute, the material demand prediction module uses a neural network
algorithm to construct a material demand prediction model, inputs the preset attribute parameter and project type attributes of a power-grid project to be predicted into the material demand prediction model, and outputs a predicted value of the material utilization amount of the corresponding power-grid project to be predicted. According to the invention, the cost of material prediction is reduced, and the prediction precision is higher.