Monday, September 27, 2010

Using pattern recognition techniques for controlling dengue epidemic in Sri Lanka

Dengue epidemic has become a threat to Sri Lanka with a fast spread all over the country. The National Dengue Prevention Unit of the Ministry of Health, Sri Lanka has reported 30,607 cases of dengue and 221 deaths during January-September, 2010 . Situation is same for many other countries as well.  The only way to get rid of this danger is the extinction of the cause-mosquitoes.

The lot of money which is spent for the prevention of this epidemic, can be optimized and maximum results can be gain if the pattern of the dengue vector-Aedes aegypti spread can be properly identified. Countries such as Singapore, USA and Cuba has used GIS (Geo Information Systems),statistical analysis and sensor network technologies to overcome this issue.

Currently, the foliole community (www.foliole.net) works on a pattern recognition technique based on fuzzy neural networks and k-means clustering. The neuro fuzzy system is used to estimate the influence of prominent factors affecting the spread of dengue vector, Aedes aegypti - precipitation, temperature, humidity and urbanization. Other socio-economic facts such as education of the people and poverty can also be considered. Meanwhile training data is prepared using k-means clustering for reported dengue cases.

Read the full article here : http://foliole.net/dengueprevention