| Resumen |
This paper introduces a novel classification model tailored explicitly for the challenges inherent in data streaming environments. Our innovative model integrates a comprehensive suite of distance measures, including the Jaccard index, Minkowski distance (both Manhattan and Euclidean variants), Pearson correlation index, and a similarity operator called Global Hybrid Online Similarity (GHOS) to enhance classification accuracy and efficiency in the dynamic nature of streaming data into the NACOD classifier. To rigorously evaluate our proposal, extensive experimentation was conducted using three extensive and imbalanced data sets representing diverse domains. Performance metrics including Balanced Accuracy, Kappa statistics and Mathews Correlation Coefficient were employed to assess the effectiveness of our model. The results of our proposal demonstrate its remarkable performance, underscoring its potential to significantly enhance classification accuracy and efficiency in real-world data streaming applications. © 2024 Copyright held by the owner/author(s). |
