Big Data Abstraction Through Multiagent Systems
Agent mining interaction has attracted a lot of attention among researchers. It is possible to solve large data mining problems through multiagent systems. Big data is characterized by huge volumes of data that are not easily amenable for generating abstr
- PDF / 246,201 Bytes
- 11 Pages / 441 x 666 pts Page_size
- 33 Downloads / 274 Views
Big Data Abstraction Through Multiagent Systems
8.1 Introduction Big Data is proving to be a new paradigm after data mining in large or massive data analytics. With increasing ability to store large volumes of data at every second, the need for making sense of the data for summarization and business exploitation is steadily increasing. The data is emanating from customer records, pervasive sensors, sense of keeping every data item for potential subsequent analysis, security paranoia, etc. Big Data theme is gaining importance especially because large volumes of data in variety of formats are found related and need to be processed in conjunction with each other. Large databases, which are conventionally built on predefined schema, are not directly usable. However, there are arguments in the literature for and against the use of Map-Reduce algorithm as compared to massive parallel databases. Such databases are built by many commercial players. Agent-mining interaction is gaining importance in research community in solving massive data problems in divide-and-conquer manner. The interaction is mutual such as agent driving data mining and vice versa. We discuss these issues in more detail in the chapter. We propose to solve Big Data analytics problems through multiagent systems. We propose few problem solving schemes. In Sect. 8.2, we provide an overview of Big Data and challenges it offers to research community. Section 8.3 discusses large data problems as solved by conventional systems. Section 8.4 contains a discussion on overlap between big data and data mining. A discussion on multiagent systems is provided in Sect. 8.5. Section 8.6 contains proposed multiagent systems for abstraction generation with Big Data.
8.2 Big Data Big data is marked by voluminous heterogeneous datasets that need to be accessed and processed in real time to generate abstraction. Such an abstraction is valuable T. Ravindra Babu et al., Compression Schemes for Mining Large Datasets, Advances in Computer Vision and Pattern Recognition, DOI 10.1007/978-1-4471-5607-9_8, © Springer-Verlag London 2013
173
174
8 Big Data Abstraction Through Multiagent Systems
for scientific or business decisions depending on nature of data. These attributes are conventionally termed as three v’s, known as volume, velocity, and variety. Some experts add an additional v, known as value. The big data analytics has also led to a new inter-disciplinary topic, called data science, which combines statistics, machine learning, natural language processing, visualization, and data mining. Associated terminologies to data science are data products and data services. The need for Big Data analytics or abstraction arose due to increasing ability to sense and store the data, omnipresence of data, ability to see the business potential of such data sets. Some examples are the trails of data that one leaves as one browses web pages, tweets his/her opinions, social media channels, visits to multiple stores to purchase varieties of items, scientific data such as genome sequenci
Data Loading...
