Establishment of Effectiveness Evaluation Indicator System based on 1D-CNN

Towards the decisive influence of substantial subjective factors in constructing effectiveness evaluation indicators, this article introduces a parallel one-dimensional convolutional neural network based on sample data. Then an objective method of constructing effectiveness evaluation indicator system is firstly put forward, including utilizing the multi-dimension convolution kernels to extract the raw data characteristics and concatenating a new indicator system. The validity and observability of the new indicator system based on sample data are verified by the simulation of indicator features extraction and effectiveness evaluation.