An efficient approach for opinion mining from skewed twitter corpus using over sampled imbalance data learning

In this paper, we have proposed a novel algorithmic approach-Over Sampled Imbalance Data Learning (OSIDL) to discover the knowledge from real world imbalanced datasets. The process of confiscate is implemented in the majority subset to remove the noisy, border line and missing instances and substitute of missing instances is done in the minority subset for improving the strength of the dataset. Experimental comparisons are done on real world twitter corpus with bench mark traditional C4.5 algorithm. Nine evaluation metrics are used for efficient opinion mining. The results suggest that the proposed OSIDL algorithm performed better than the compared algorithm in terms of Accuracy, AUC, Precision, F-score, Recall, TP Rate and TN Rate. The experimental results also clearly indicate the effectiveness of the proposed approach in the scenario of skewed twitter corpus.