Artificial intelligence to assist specialists in the detection of haematological diseases

• DOI: 10.1016/j.heliyon.2023.e15940
• Publication Date: 2023-05-03T19:46:56Z
• Authors: Sergio Diaz-del-Pino; Roberto Trelles-M...; F.A. González-Fer...; Nicolas Guil
• Keywords: Interpretability; Biomedicine; Gradient boosting

Artificial intelligence, particularly the growth of neural network research and development, has become an invaluable tool for data analysis, offering unrivalled solutions image generation, natural language processing, personalised suggestions. In meantime, biomedicine been presented as one pressing challenges 21st century. The inversion age pyramid, increase in longevity, negative environment due to pollution bad habits population have led a necessity methodologies that can help mitigate fight against these changes.The combination both fields already achieved remarkable results drug discovery, cancer prediction or gene activation. However, such labelling, architecture improvements, interpretability models translational implementation proposals still remain. haematology, conventional protocols follow stepwise approach includes several tests doctor-patient interactions make diagnosis. This procedure significant costs workload hospitals.In this paper, we present artificial intelligence model based on networks support practitioners identification different haematological diseases using only rutinary inexpensive blood count tests. particular, binary multiclass classification specialised where is studied combined along it, taking into account clinical knowledge problem, obtaining up 96% accuracy experiment. Furthermore, compare method traditional machine learning algorithms gradient boosting decision trees transformers tabular data. use techniques could reduce cost time improve quality life specialists patients while producing more precise diagnoses.