Research on a Classification Prediction Model for Schizophrenic Patients Based on PH-GAT
This paper studies the current analysis based on cerebral network,the study shows that the analysis methods can be broadly categorized into two main approaches:analysis based on continuous homotopy methods and analysis based on Deep Learning models.In order to enhance the predictive capabilities of brain disease diagnosis,this model incorporates continuous homotopy into the GAT model,endowing it with a"topological awareness".Towards the end of the model,the Long Short-Term Memory(LSTM)model is employed to capture temporal information embedded within the extracted features,thereby enhancing the effectiveness of classification prediction.Under the PH-GAT model,a fusion of local and global features is applied for classifying data in the Theta frequency range,achieving a high classification accuracy of 0.930 9.This approach not only enables the discovery of objective and effective imaging biomarkers for early schizophrenia diagnosis,but also enhances the predictive capabilities of brain disease diagnosis.
国家科技期刊平台
NSTL
万方数据
维普
