Analysis Classification Imbalanced Data Predection Multiclass in Machine Learning: Oversampling with SMOTE and ADASYN

Imbalanced datasets are prevalent in various machine learning applications, especially in financial credit prediction data, where most of the examples represent financial credit. This research addresses the challenge of class imbalance in multiclass financial credit datasets by modifying based on multiclass logistic regression. SMOTE and ADASYN methods are used for the resampling process. The resampling technique is used to preserve the original data. Performance evaluation is performed on financial credit data. Comprehensive analysis was conducted with previous research using the same data set. Test results using accuracy, recall, fi-score and auc-roc testing are attached in table 4.