Open

Fulvio Favilla Filho is an Electrical Engineering Bachelor student at the University of São Paulo in Brazil.  Fulvio has been developing a study related to cardiovascular disease prediction using Machine Learning models as his bachelor's dissertation. The HPCC Systems platform was used to perform a comparative analysis between different ML algorithms to evaluate which model is more accurate to predict the presence or absence of cardiovascular diseases in patients by using simple clinical data as input.

Poster Abstract

Cardiovascular diseases (CVD) are the leading cause of global mortality, presenting a significant challenge to public health. The ability to accurately predict the presence of CVD in patients holds immense potential for improving treatment strategies and enhancing clinical decision-making. To address this crucial issue, the present research proposes an approach that utilizes machine learning techniques for CVD 
prediction.

The main objective of this study is to create predictive models through HPCC Machine Learning bundles, comparing models based on specific metrics to select the most appropriate one for prediction. Through the analysis and interpretation of simple clinical data, the model seeks to provide accurate predictions to identify patients at risk of developing CVD, enabling preventive interventions and more effective treatments.

The dataset used for model development contains patient records, encompassing various types of features: objective, examination, and subjective. In 
addition to the creation of predictive models, an analysis of the importance of each feature during CVD prediction is conducted, allowing for a better understanding of the relevance of different characteristics in cardiovascular health.

After data collection, the process of model development involves the following steps:

Data preprocessing: In this stage, the data is prepared for analysis, including outlier removal, balancing, normalization, and encoding of categorical variables.

Model training: HPCC bundles Learning Trees and Logistic Regression models are trained with the preprocessed data to predict CVD outcomes.

Evaluation and optimization: The models are evaluated based on performance metrics. An optimization of the algorithm’s hyperparameters is conducted, along with subsequent data processing used in training, to achieve the best possible results.

Through this study, the aim is to advance the field of cardiovascular disease prediction. The successful application of machine learning models in this context can have a significant impact on public health, reducing the morbidity and mortality associated with cardiovascular diseases. Additionally, the analysis of feature importance in the prediction process provides valuable insights for a deeper understanding of the factors contributing to the development of these diseases, providing a solid foundation for future research and more effective preventive interventions.

Presentation

In this Video Recording, Fulvio provides a tour and explanation of his poster content.

Machine Learning Approach to Cardiovascular Disease (CVD) Prediction

Click on the poster for a larger image.

Open