Implementation of hybrid classification model in distributed systems for network monitoring

The main objective of this study is to compare SVM (Support Vector Machine) classification algorithm and Neural Networks classification algorithms identify the pitfalls and propose a new hybrid classification algorithm which is reliable, fast, efficient and robust handling large data sets. A new version of Support Vector Machine algorithm is designed and developed in the study which is used for training the dataset followed by testing using Neural Networks classification algorithm. Generally most of the classification algorithms are working well for small and moderate data, while going for large datasets, the efficiency drops, this study analyses all these factors and proposing a new hybrid model, which solves all these drawbacks. The second main objective of this study is to analyze whether the number of back propagation steps are minimized in Neural Networks algorithm. This model can be used to monitor and predict the networking issues that occur in distributed systems.