Publicação
Predictive data mining in nutrition therapy
| Resumo: | The assessment and measurement of health status in communities throughput the world is a massive information technology challenge. Data mining, plays a vital role in health care industry since it really has the potential to generate a knowledge-rich environment that reduces medical errors, decreases costs by increasing efficiency, improves the quality of clinical decisions and significantly enhances patient's outcomes and quality of life. This study falls within the context of nutrition evaluation and its main goal is to apply classification algorithms in order to predict if a patient needs to be followed by a nutrition specialist. One of the tools resorted in this study was the Waikato Environment for Knowledge Analysis (Weka in advance) Workbench since it allows to quickly try out and compare different machine learning solutions. The tasks involved in the development of this project included data preparation, data preprocessing, data transformation and cleaning, application of several classifiers and its respective evaluation through performance measures that include the confusion matrix, accuracy, error rate, and others. The accomplished results showed to be quite optimistic presenting promising values of performance measures. specifically an accuracy around 91 %. |
|---|---|
| Autores principais: | Ferreira, Diana |
| Outros Autores: | Peixoto, Hugo; Machado, José Manuel; Abelha, António |
| Assunto: | Classification algorithms Clinical decisions Data mining Health care Information technology Machine learning Nutrition evaluation Performance measures |
| Ano: | 2018 |
| País: | Portugal |
| Tipo de documento: | comunicação em conferência |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | The assessment and measurement of health status in communities throughput the world is a massive information technology challenge. Data mining, plays a vital role in health care industry since it really has the potential to generate a knowledge-rich environment that reduces medical errors, decreases costs by increasing efficiency, improves the quality of clinical decisions and significantly enhances patient's outcomes and quality of life. This study falls within the context of nutrition evaluation and its main goal is to apply classification algorithms in order to predict if a patient needs to be followed by a nutrition specialist. One of the tools resorted in this study was the Waikato Environment for Knowledge Analysis (Weka in advance) Workbench since it allows to quickly try out and compare different machine learning solutions. The tasks involved in the development of this project included data preparation, data preprocessing, data transformation and cleaning, application of several classifiers and its respective evaluation through performance measures that include the confusion matrix, accuracy, error rate, and others. The accomplished results showed to be quite optimistic presenting promising values of performance measures. specifically an accuracy around 91 %. |
|---|