Averaging GNAR forecasts across the top five random graphs selected by recent one-step-ahead errors beats AR benchmarks at all horizons and beats the Bank of England at 4-6 months, though the Bank comparison lacks significance testing.
The weighted random graph model
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We introduce the weighted random graph (WRG) model, which represents the weighted counterpart of the Erdos-Renyi random graph and provides fundamental insights into more complicated weighted networks. We find analytically that the WRG is characterized by a geometric weight distribution, a binomial degree distribution and a negative binomial strength distribution. We also characterize exactly the percolation phase transitions associated with edge removal and with the appearance of weighted subgraphs of any order and intensity. We find that even this completely null model displays a percolation behavior similar to what observed in real weighted networks, implying that edge removal cannot be used to detect community structure empirically. By contrast, the analysis of clustering successfully reveals different patterns between the WRG and real networks.
fields
stat.AP 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Forecasting UK Consumer Price Inflation with RaGNAR: Random Generalised Network Autoregressive Processes
Averaging GNAR forecasts across the top five random graphs selected by recent one-step-ahead errors beats AR benchmarks at all horizons and beats the Bank of England at 4-6 months, though the Bank comparison lacks significance testing.