Global public health continues to face substantial obstacles from mosquito-borne diseases, making precise and effective techniques for mosquito species identification necessary. We present a unique method in this article called "Mosquito Species Classification through Wingbeat Analysis: A Hybrid Machine Learning Approach," which uses wingbeat analysis and deep learning techniques to classify mosquito species. Our approach leverages Convolutional Neural Networks (CNNs) as the core model to provide robust and dependable classification performance.

We make use of an extensive dataset that includes wingbeat recordings from many species of mosquitoes and apply comprehensive preprocessing and feature engineering techniques to enhance the model's effectiveness. Specifically, we extract and combine features such as zero crossing rate (ZCR), root mean square energy (RMSE), mel-frequency cepstral coefficients (MFCC), as well as augmented features derived from audio transformations like add_noise, shifting, pitching, and stretching. This combination of handcrafted and augmented features helps to enrich the training data and improve the generalizability of the model.

After thorough testing and evaluation, we demonstrate that our CNN-based method achieves superior performance in accurately classifying various mosquito species. Our findings underscore the potential of deep learning methods, particularly CNNs, to surpass conventional classification techniques in species identification tasks. Additionally, we highlight the critical role of accurate species classification in vector surveillance and epidemiological research, emphasizing the broader impact of our work on ecological studies and disease control strategies.

Keywords: Deep learning, CNN, species classification, wingbeat analysis, mosquito-borne diseases, ZCR, RMSE, MFCC, data augmentation.