Real estate valuation method and system based on neural network statistic model

The present invention provides a real estate valuation system based on a neural network statistic model. The system comprises a training unit, a neural network unit and an evaluation unit. The evaluation unit quantizes the grade of the influence factor of a to-be-estimated property so as to obtain the quantized value of the influence factor of the to-be-estimated property according to a quantization standard. The training unit stores the preset parameters of each neural network. In each neural network, the quantized value of the influence factor of the to-be-estimated property is adopted as an input variable, so that the evaluated value of each neural network for the to-be-estimated property is obtained. The evaluation unit multiplies the value of the to-be-estimated property evaluated by each neural network by a preset weight value so as to obtain the final evaluated value of the to-be-estimated property. According to the technical scheme of the invention, neural networks are determined according to administ