Temporal effects in trend prediction: identifying the most popular nodes in the future
arXiv:1412.6753 · doi:10.1371/journal.pone.0120735
Abstract
Prediction is an important problem in different science domains. In this paper, we focus on trend prediction in complex networks, i.e. to identify the most popular nodes in the future. Due to the preferential attachment mechanism in real systems, nodes' recent degree and cumulative degree have been successfully applied to design trend prediction methods. Here we took into account more detailed information about the network evolution and proposed a temporal-based predictor (TBP). The TBP predicts the future trend by the node strength in the weighted network with the link weight equal to its exponential aging. Three data sets with time information are used to test the performance of the new method. We find that TBP have high general accuracy in predicting the future most popular nodes. More importantly, it can identify many potential objects with low popularity in the past but high popularity in the future. The effect of the decay speed in the exponential aging on the results is discussed in detail.
References in corpus (10)
- Emergence of scaling in random networks
- Quantifying Long-Term Scientific Impact
- Robust dynamic classes revealed by measuring the response function of a social system
- Characterizing and modeling the dynamics of online popularity
- The first-mover advantage in scientific publication
- Temporal effects in the growth of networks
- The effect of aging on network structure
- Prediction of highly cited papers
- Trend prediction in temporal bipartite networks: the case of Movielens, Netflix, and Digg
- Harvesting Collective Intelligence: Temporal Behavior in Yahoo Answers