Navigation mit Access Keys

Menu principal

 
 

Predictive Analytics in Big Data

 

Predictive analytics deals with extracting patterns from series of data, such as observational, in order to better understand and predict the evolvement of spatio-temporal phenomena.

Big data refers to data of varying type and quality that are produced with high velocity and in high volumes such that they cannot be processed by traditional methods.

The objective of the project is to develop innovative methods and models in order to assess the risk of an outbreak of a contagious disease and to predict the spatial and temporal patterns of disease spread such that decision makers can be supported on taking timely measures.

 

 

Predictive analytics deals with extracting patterns from series of data, such as observational, in order to better understand and predict the evolvement of spatio-temporal phenomena.

 

Big data refers to data of varying type and quality that are produced with high velocity and in high volumes such that they cannot be processed by traditional methods.

 

The objective of the project is to develop innovative methods and models in order to assess the risk of an outbreak of a contagious disease and to predict the spatial and temporal patterns of disease spread such that decision makers can be supported on taking timely measures.

 

Predictive analytics deals with extracting patterns from series of data, such as observational, in order to better understand and predict the evolvement of spatio-temporal phenomena.

 

Big data refers to data of varying type and quality that are produced with high velocity and in high volumes such that they cannot be processed by traditional methods.

The objective of the project is to develop innovative methods and models in order to assess the risk of an outbreak of a contagious disease and to predict the spatial and temporal patterns of disease spread such that decision makers can be supported on taking timely measures. The project is an integral part of the NRP 75 project "PIG DATA: Health Analytics for the Swiss Swine Industry". Particularly the case of contagious diseases in pigs potentially affecting also humans and other species will be studied. The developed methods and models are expected to be applicable across species and topical boundaries, for instance, to predict emerging phenomena in ecosystem research.

The figure shows where the project starts from by taking the example of avian influenza: Prior to the departure of migrating birds in autumn 2006, the Swiss authorities established zones of 1 kilometer along the waterfront of some major rivers and lakes of the Swiss Plateau in which poultry had to be kept in the henhouse. The reasons behind were that cases of human- (and swine-) pathogenic H5N1 had been found primarily in water birds on the large lakes and the authorities suspected migrating birds to transmit the virus.

This measure was based on a simple spatio-temporal predictive model with input variables "exposed waterbody" and "season" and output variable "protection zone". The development of more complex models and of methods that are more sophisticated than computing a buffer around exposed waterbodies is envisaged in the project presented here.