Prediction of dengue outbreak in Selangor Malaysia using machine learning techniques
Nurul Azam Mohd Salim, Yap Bee Wah, Caitlynn Reeves, Madison Smith, Wan Fairos Wan Yaacob, Rose Nani Mudin, Rahmat Dapari, Nik Nur Fatin Fatihah Sapri et al.
importance 7/10ai assistedMosquito-borne diseases and controlpublic healthstage: research
Problem
Predicting dengue fever outbreaks in five high-incidence districts of Selangor, Malaysia using climate variables.
Why it matters
Early outbreak prediction would let health officials implement timely preventative actions in an endemic region.
What the AI did
Machine learning models using temperature, wind speed, humidity and rainfall were trained to predict dengue outbreaks.
Results
Best model: Accuracy 70%, Sensitivity 14%, Specificity 95%, Precision 56%; sensitivity rose to 63.54% on testing sample.
Impact
Peer-reviewed but modest performance; low sensitivity means most outbreaks are missed.
Record
Venue
Scientific Reports
Published
2021
Work type
article
Authors
Nurul Azam Mohd Salim, Yap Bee Wah, Caitlynn Reeves, Madison Smith, Wan Fairos Wan Yaacob, Rose Nani Mudin, Rahmat Dapari, Nik Nur Fatin Fatihah Sapri et al.
dengue, outbreak prediction, machine learning, SVM, climate variables, Malaysia
Importance rating
7/ 10
significant — thousands to millions affected
Dengue is a major mosquito-borne disease; abstract states it affects nearly 3.9 billion people globally and is endemic in Malaysia, so outbreak prediction has large public health value.
Basis: stated.
The source itself states the scale.
Stated scale: nearly 3.9 billion people globally affected by dengue; endemic in Malaysia, highest concentration in Selangor. Method and limits.