Design and Analysis of a Prediction System about Influenza-Like Illness from the Latent Temporal and Spatial Information

Xiaoxiang Guo, Neal N. Xiong, Haiyan Wang, Jingli Ren

Research output: Contribution to journalArticlepeer-review

5 Scopus citations


Influenza poses a significant risk to public health, as evidenced by the 2009 H1N1 pandemic which caused up to 203 000 deaths worldwide. Predicting the spatiotemporal information of disease in the incubation period is crucial because the prime aim of it is to provide guidance on preparing a response and avoid presumably adverse impact caused by a pandemic. This article designs and analyzes a prediction system about influenza-like illness (ILI) from the latent temporal and spatial information. In this system 1) Gaussian function model and multivariate polynomial regression are employed to investigate the temporal and spatial distribution of ILI data; 2) the phase space reconstructed by delay-coordinate embedding is used to explore the dynamical evolution behavior of the 1-D ILI series; and 3) a dynamical radial basis function neural network (DRBFNN) method which is the kernel of the system, is proposed to predict the ILI values based on the correlations between the observations space and reconstructed phase space. The performance analysis of our system shows that the regression equations coupling with spatial distribution information can be used to supplement the missing data, and the proposed DRBFNN method can predict the trends of ILI for the following one year. Furthermore, the prediction system in this article applies a model-free control schemes, i.e., there are no restriction equations between the multivariable inputs and outputs. This prediction system is expected to be used in predicting the output signals, even the chaotic output signals, in meteorology, industry, medicine, economy, and other fields. An example of predicting the Standard Poors 500 index is given to introduce the application of our proposed system. The trend of open prices of the following eight trading days is well predicted.

Original languageEnglish (US)
Pages (from-to)66-77
Number of pages12
JournalIEEE Transactions on Systems, Man, and Cybernetics: Systems
Issue number1
StatePublished - Jan 1 2022


  • Dynamical radial basis function neural network (DRBFNN)
  • influenza-like illness (ILI)
  • multivariate regression

ASJC Scopus subject areas

  • Software
  • Control and Systems Engineering
  • Human-Computer Interaction
  • Computer Science Applications
  • Electrical and Electronic Engineering


Dive into the research topics of 'Design and Analysis of a Prediction System about Influenza-Like Illness from the Latent Temporal and Spatial Information'. Together they form a unique fingerprint.

Cite this