Active Learning from an Imbalanced Dataset: A Study Conducted on the Depression, Anxiety, and Stress Dataset
- Title
- Active Learning from an Imbalanced Dataset: A Study Conducted on the Depression, Anxiety, and Stress Dataset
- Creator
- Umme Salma M.; Amala Ann K.A.
- Description
- The proposed chapter deals with psychological data related to depression, anxiety, and stress to study how the classification and analysis is carried out on imbalanced data. The proposed study not only contributes on providing practical information about the balancing techniques such as synthetic minority oversampling technique but also reveals the strategy for dealing with the working of many existing classification algorithms such as the support vector machine, random forest, XGBoost, etc. on the imbalanced dataset. The present use of evaluation metrics that are solely implied for the imbalanced data classification is also illustrated. It was observed that the ordinary model assessment techniques do not precisely quantify model execution when gone up against imbalanced datasets and that the common techniques such as the logistic regression and decision tree have a predisposition toward classes that have many observations. The attributes of the minority class are treated low and are routinely overlooked. Henceforth, there is a high likelihood of misclassification of the minority class when compared to the majority class. A confusion matrix which contains data about the real and predicted class is used as an assessment standard to check the exhibition of grouping calculation. Rather than going for accuracy, F-score and the area under the curve are considered as the measures to evaluate the classification model. 2022 selection and editorial matter, Vishal Jain, Sapna Juneja, Abhinav Juneja, and Ramani Kannan.
- Source
- Handbook of Machine Learning for Computational Optimization: Applications and Case Studies, pp. 251-266.
- Date
- 2021-01-01
- Publisher
- CRC Press
- Coverage
- Umme Salma M., Department of Computer Science, CHRIST (Deemed to be University), Bangalore, India; Amala Ann K.A., Data Science Department, CHRIST (Deemed to be University), Bangalore, India
- Rights
- Restricted Access
- Relation
- ISBN: 978-100045567-0; 978-036768542-3
- Format
- Online
- Language
- English
- Type
- Book chapter
Collection
Citation
Umme Salma M.; Amala Ann K.A., “Active Learning from an Imbalanced Dataset: A Study Conducted on the Depression, Anxiety, and Stress Dataset,” CHRIST (Deemed To Be University) Institutional Repository, accessed February 24, 2025, https://archives.christuniversity.in/items/show/18747.