phat 24(1):

Research Article

Prediction of Anemia using various Ensemble Learning and Boosting Techniques

Download138 downloads
  • @ARTICLE{10.4108/eetpht.9.4197,
        author={Nalluri Shweta and Sagar Dhanraj Pande},
        title={Prediction of Anemia using various Ensemble Learning and Boosting Techniques},
        journal={EAI Endorsed Transactions on Pervasive Health and Technology},
        volume={9},
        number={1},
        publisher={EAI},
        journal_a={PHAT},
        year={2023},
        month={10},
        keywords={Anemia, Prediction, Machine Learning, Random Forest, Ensemble learning, Boosting},
        doi={10.4108/eetpht.9.4197}
    }
    
  • Nalluri Shweta
    Sagar Dhanraj Pande
    Year: 2023
    Prediction of Anemia using various Ensemble Learning and Boosting Techniques
    PHAT
    EAI
    DOI: 10.4108/eetpht.9.4197
Nalluri Shweta1, Sagar Dhanraj Pande1,*
  • 1: Vellore Institute of Technology University
*Contact email: sagarpande30@gmail.com

Abstract

INTRODUCTION: Anemia is a disease of great concern. It is mainly seen in people who are deficient in several vitamins like B12 and those who are deficient in iron. Neglecting the situation and leaving it untreated could lead to severe consequences in the future. Hence it is of great importance to predict Anemia in an individual and treat it in the optimum stage. OBJECTIVES: In this paper, machine learning was used for the prediction of Anemia. METHODS: The dataset used for this was formed by combining different datasets from Kaggle. The accuracy of various machine learning techniques was evaluated to find out the best one. Along with the supervised learning algorithms like Random Forest, SVM, Naive Bayes etc., Linear Discriminant Analysis, Quadratic Discriminant Analysis and ensemble learning methods were also performed. RESULTS: Upon evaluation, among the best performers, the execution time was also taken into consideration to determine which classifier works well. Among all the algorithms used, XGboost worked the best with an optimum execution time. CONCLUSION: The conclusion is that for the data used in the work, XGboost results as the best model.