Unsupervised dimension reduction for data clustering and improving signal-to-noise-ratio
Recently, machine learning has been used in every possible field to leverage its amazing power. In this project we employ and advance machine learning algorithms for analysis of networks log data due to extraction of informative features. These data, which are recorded every millisecond, are usually high dimensional and imbalanced where no class label is assigned to them. We propose to realize data analysis through simultaneous performing of dimensionality reduction and data clustering incorporating local characteristics of the sample space to handle data imbalancity and variations. The proposed approaches for dimension reduction and clustering will be great assets in design, optimization and operations of the partner organization operational system. They also can improve the signal-to-noise-ratio and anomaly detection.