Firefly based feature selection algorithms for big data classification

Abstract

In general, the big data is defined by three major characteristics like volume, variety as well as velocity. This directly means that in some point of time when the variety, volume, and velocity of data get increased, the available technologies and techniques might not have the ability to handle both data storage and data processing as well. In the research work, Big Data Analytics is actually the progression of data characteristic analysis and understanding of immense size datasets via extracting or mining valuable geometric and statistical patterns. However, the data analytics generally faces many challenges by means of storage and data processing. It is very important to ensure the accuracy, completeness, and time maintenance during data usage for making accurate timely decisions. Data in terabytes would take more time to upload in cloud and furthermore this data would rapidly change and makes the real time uploading harder. Similarly, the distributed nature of cloud is also challenging for big data analysis. This big data analysis is the progression of applying improved analytics and visualization models for large data sets for uncovering hidden patterns and unknown correlations that gains efficient decision making or classification. The designing must be focussed in terms of different performance criteria that are crucial in data classification techniques. (i) Variety as well as Heterogeneous Data is gathered from different sources, which generate data. (ii) Stability: High stability of mining tools and data management is needed to handle the big data. (iii) Speed/Velocity: The data should be rapidly evaluated and processed else the outcomes obtained from the data would become worthless. (iv) Trust and Accuracy: More data sources are there from where data is gathered, and that might not be trustable or verifiable. (v) Privacy is also a significant issue in data analytics. Finally, (vi) Inter-activeness is the major complexity since it permits the users in visualizing, evaluating and interpreting intermediate and last results. newline