INTERNATIONAL JOURNAL OF SCIENTIFIC DEVELOPMENT AND RESEARCH International Peer Reviewed & Refereed Journals, Open Access Journal ISSN Approved Journal No: 2455-2631 | Impact factor: 8.15 | ESTD Year: 2016
open access , Peer-reviewed, and Refereed Journals, Impact factor 8.15
One of the most extensively used modelling strategies in the world of machine learning is classification algorithms like K- Nearest Neighbor (KNN) and DECISION TREE Algorithm. Machine learning models using classification algorithms are widely utilized in a variety of domains, including data analytics, image classification, computer vision, exploratory analysis, and game Artificial Intelligence, among others. In this case, the model must be extremely accurate, versatile, and efficient in order to successfully complete the work at hand. However, regardless of technique, the metrics used to assess a model's effectiveness are influenced by a variety of elements such as the confusion matrix, accuracy score, and so on. Among all these factors, the balance between bias and variance must be carefully maintained to optimize the model's performance. The KNN and DECISION TREE algorithms will be used to investigate bias and variance on the Maternal Health Risk Dataset in this research. Furthermore, the paper will focus on the regularization process and its impact on the balance of bias and variance, as well as how to deal with any inconsistencies that may develop owing to minor changes in dataset values. The benefits and drawbacks of variable bias and variance values, respectively, indicate the level of model adaptability on a dataset, regardless of how the training, testing, and validation data are divided.
Keywords:
K-NN, Decision Tree, Variance, Machine Learning
Cite Article:
"Bias variance trade off on maternal health risk dataset in K-NN and Decision tree algorithm", International Journal of Science & Engineering Development Research (www.ijsdr.org), ISSN:2455-2631, Vol.7, Issue 11, page no.18 - 22, November-2022, Available :http://www.ijsdr.org/papers/IJSDR2211004.pdf
Downloads:
000337068
Publication Details:
Published Paper ID: IJSDR2211004
Registration ID:202424
Published In: Volume 7 Issue 11, November-2022
DOI (Digital Object Identifier):
Page No: 18 - 22
Publisher: IJSDR | www.ijsdr.org
ISSN Number: 2455-2631
Facebook Twitter Instagram LinkedIn