Network Graphs for Epidemiology

SEIR model

import networkx as nx
import ndlib.models.ModelConfig as mc
import ndlib.models.epidemics as ep
from ndlib.viz.mpl.DiffusionTrend import DiffusionTrend

# Network topology
g = nx.erdos_renyi_graph(1000, 0.1)                    # 1000 nodes with 10% probability of edge creation

# Model selection
model  = ep.SEIRModel(g)
model2 = ep.SEIRctModel(g)
model3 = ep.SEISModel(g)

# Model Configuration
cfg = mc.Configuration()
cfg.add_model_parameter('beta', 0.01)                   # Infection probability     S -> I
cfg.add_model_parameter('lambda', 0.9)                  # Recovery probability      I -> S
cfg.add_model_parameter('gamma', 0.005)                 # Removal probability       I -> R
cfg.add_model_parameter('alpha', 0.05)                  # Latent Period             E -> I
cfg.add_model_parameter("fraction_infected", 0.05)      # 50 infected nodes at start in the network 
model.set_initial_status(cfg)
model2.set_initial_status(cfg)
model3.set_initial_status(cfg)

# Simulation execution
iterations = model.iteration_bunch(500)
trends = model.build_trends(iterations)
trends2 = model2.build_trends(iterations)
trends3 = model3.build_trends(iterations)

# Show simulation
viz = DiffusionTrend(model, trends)
viz.plot("diffusion")