الفهرس | Only 14 pages are availabe for public view |
Abstract Big Data analysis play a very important role in many real-world applications, especially, biomedical data sets. Unfortunately, the majority of these data are imbalanced because, there is a small number of instances in one class (minority class) compared to the other class (majority class). The standard classifiers are not capable of handling such data sets. Hence, a new technique for dealing with such data sets is required. For solving local optima problem in high dimensional data sets, this thesis proposes an imbalanced biomedical big data deep learning framework (WOA+BRNN) that consists of three developed phases. The first phase is the feature selection phase, which uses the WOA (Whale optimization algorithm) for finding the best set of features. The second phase is the preprocessing phase, which uses the SMOTE algorithm and the LSH-SMOTE algorithm for solving the class imbalance problem. Lastly, the third phase is WOA+BRNN framework, which is using the Whale optimization algorithm for training a deep learning approach called Bidirectional Recurrent Neural Network (BRNN). The proposed framework WOA+BRNN has been tested against 9 highly imbalanced data sets one of them is big data set in terms of AUC (Area Under Curve) against four of the most common use machine-learning algorithms (Naïve Bayes, AdaBoostM1, Decision Table, Random Tree), in addition to GWO-MLP (Training Multi-Layer Perceptron using Greywolf optimizer), then we test our algorithm over 4 well-known data sets against GWO-MLP and Particle Swarm Optimization (PSO-MLP), Genetic Algorithm (GA-MLP), Ant Colony Optimization (ACO-MLP), Evolution Strategy (ES-MLP), and Population-based Incremental Learning (PBIL-MLP) in terms of classification accuracy. Experimental results proved that our proposed algorithm WOA+BRNN has achieve promising accuracy and high local optima avoidance, and outperformed four of the most common use machine-learning algorithms, and GWO-MLP in terms of AUC (Area Under Curve). |