special-topic-cse

WI4450 - Special Topics in Computational Science and Engineering.

Implementation of a project were we focus our research on the main two measures presented by O’Brien et al.: The entropy and Hierarchy of a network. We start of our study by investigating the relationship between the structure of networks and their non-normality level. Furthermore, we have extended our research including two additions: Robustness analysis of generated datasets, and analysis of the entropy via Quantum walks.

Report: to be published.

Install

pip install -r requirements.txt

Price's model

An implementation of Price's can be found in /models/prices_model.py. The model accepts 5 parameters:

• n the number of nodes in the generated graph.
• m the number of outgoing edges that a new node has.
• k0 a parameter that ensures that citations are made. Usually k0=.
• n0 the number of initial nodes in the graph.
• reciprocal_threshold or p the reciprocal threshold.

Dataset

The dataset that is used in for the report can be found in the /dataset folder. The dataset is generated by generate_dataset.py. Information about the dataset and the computed measures for each graphs can be found in the prices_model_seed1313_m3_k01_initial_nodes1.csv file.

Perturbations of the dataset

Perturbations are performed to test the robutsness of the graph in the dataset. 4 types of perturbations are implemented: node removal, edge addition, edge removal and edge reversal. The file perturbation_loop.py contains code to runs these perturbations on the whole dataset and stores all the measures of the perturbated graphs in the /output folder. All the resulting figures can be found in /plots. Also a parallel implementation is added in perurbation_loop_parallel.py.

Quantum random walk

The implementation for the quantum random walk can be found in /quantum-walk.