Student Research Reports

View Research Report

Anopheles mosquitoes are densely populated in Africa but can be found in other regions and continents, including but not limited to North America, South America, Europe, and Asia (specifically the Indian subcontinent). A point of concern with Anopheles mosquitoes is their unique ability to carry and transmit malaria, a parasitic infection that attacks red blood cells. WHO studies report that this deadly disease infects more than 200 million and kills over 500,000 people annually. The severity of this disease illustrates the need to engineer a practical method to mitigate the spread. The GLOBE Mosquito Habitat Mappers (MHM) app allows users worldwide to photograph and identify mosquito larva they encounter or trap. The collected data is available in a global database, accessible to researchers for labeling and classifying, a key step in tracking the population growth of the Anopheles mosquitoes. However, the manual verification for the classification of this data is time-consuming and inefficient, limiting the expansion of mosquito research. The use of Machine Learning (ML), a subset of Artificial Intelligence (AI), has substantial benefits for practical implementation, serving to improve image classification of mosquito larva of the Anopheles genus. This project proposes a method of implementing logistic regression for developing a mosquito identification model, with an accuracy greater than 80%, and plotting detected locations on a global map. A total of 3,275 images were extracted from the GLOBE MHM application. Each image is classified based on the absence of a siphon, a distinctive feature of the Anopheles mosquito, allowing accurate identification of the genus. The publicly available machine learning model and precise mapping of detected mosquito locations on a comprehensive world map will help mosquito ecologists, governments, and public health organizations effectively track and mitigate the spread of mosquito-borne diseases. Index Terms – Artificial Intelligence, Data Classification, Logistic Regression, Machine Learning.