Developing Deep Learning Models For Analyzing And Understanding Complex Graph Structures

Abstract

In this research, we construct deep learning (DL) networks to comprehend and analyze difficult graph structures. Due to DL’s remarkable rise in popularity in recent years, difficult tasks in difficult domains like visual analysis and linguistic processing are resolved with great efficiency. In light of this achievement, the deep neural network (DNN) replicates lower-level neural processes. It is challenging to describe such visualizations and they aren’t particularly helpful for comprehending the decision-making process. By incorporating Recurrent Neural Network (RNN) in co-activation graphs and analyzing their capacity to improve understanding, we develop prior investigations in this work. To better understand the connection among sets of neuron layers from hidden layers and output categories, the co-activation graph accumulates relationships among the activation levels of the neurons. This research is expanded to take into account various data sets that were gathered from public sources to validate the validity of the findings. By examining behaviors and strong nodes and utilizing graph imaging techniques, this work showed an understanding of the RNN function and the understanding of process modeling requirements.