Anomaly Detection Framework for Big Data

Financial service organizations have a number of strategic goals including the acquisition and retention of new and existing customers through the application of various management methodologies. In view of these goals, the institutions generate large volumes of data on profile data, purchase and browsing history and social media data daily. In recent times, businesses have increased the volume of data they process and analyze and this comes at a high computational cost in its analysis. Due to the high increase in the volume of data, most institutions are moving away from the measure of the data volume by volume and considering other parameters. The additional parameters of interest to support evaluation of systems-enabled strategy include variety and velocity. My motivated to design and employ techniques that will stem from both approaches to designing a more hybrid approach which will perform better and also moderate in terms of computational cost.