BULERIA. Repositorio Institucional de la Universidad de León
Publication
Artículo científico (JournalArticle)
Enhanced prediction of spine surgery outcomes using advanced machine learning techniques and oversampling methods
Publicación en abierto financiada por el Consorcio de Bibliotecas Universitarias de Castilla y León (BUCLE), con cargo al Programa Operativo 2014ES16RFOP009 FEDER 2014-2020 DE CASTILLA Y LEÓN, Actuación:20007-CL - Apoyo Consorcio BUCLE, [EN] Purpose: Accurate prediction of spine surgery outcomes is essential for optimizing treatment strategies. This study
presents an enhanced machine learning approach to classify and predict the success of spine surgeries, incorporating
advanced oversampling techniques and grid search optimization to improve model performance.
Methods: Various machine learning models, including GaussianNB, ComplementNB, KNN, Decision Tree, KNN with
RandomOverSampler, KNN with SMOTE, and grid-searched optimized versions of KNN and Decision Tree, were
applied to a dataset of 244 spine surgery patients. The dataset, comprising pre-surgical, psychometric, socioeconomic,
and analytical variables, was analyzed to determine the most efficient predictive model. The study explored the
impact of different variable groupings and oversampling techniques.
Results: Experimental results indicate that the KNN model, especially when enhanced with RandomOverSampler
and SMOTE, demonstrated superior performance, achieving accuracy values as high as 76% and an F1-score of 67%.
Grid-searched optimized versions of KNN and Decision Tree also yielded significant improvements in predictive accuracy
and F1-score.
Conclusions: The study highlights the potential of advanced machine learning techniques and oversampling
methods in predicting spine surgery outcomes. The results underscore the importance of careful variable selection
and model optimization to achieve optimal performance. This system holds promise as a tool to assist healthcare
professionals in decision-making, thereby enhancing spine surgery outcomes. Future research should focus on further
refining these models and exploring their application across larger datasets and diverse clinical settings., SI