A SYSTEM FOR IMBALANCE DATA CLASSIFICATION AND A METHOD THEREOF

A system of imbalance data classification, the system comprises ofa classification module connected to the system, wherein the classification module comprises of a local distance learning module applied on a nearest neighbor for classification of an imbalance dataset, a multiple distance metric module connected to the local learning algorithm to investigate the dataset, and a decision-making module associated with the distance metric module for calculating distance among the data and dataset, wherein the decision is taken using an Adadelta rule. Dataset Samples Dimensions Classes Iris 150 4 3 Breast cancer 685 9 2 Wine 178 13 3 Diabetes 768 8 2 Glass 214 9 6 E-coli 336 7 5 Yeast 1484 8 3 Predicted positive Predicted negative Actual positive True positive False positive Actual negative False negative True negative Dataset Accuracy Precision Recall F-mneasure G-mean AUC Iris 97.36 98 97 97 97.36 90.62 Breast cancer 92.30 93 92 92 90.82 91.064 Wine 71.11 72 71 71 71 67.35 Diabetes 76.04 76 76 75 69.11 71.38 Glass 70.37 67 70 68 73.31 100 E-coli 80.95 80 81 79 88.78 83.64 Yeast 56.87 57 57 55 69.58 50.0