Data Stream Mining using Reinforcement Learning: A Survey
Journal Title: International Journal of Research in Computer and Communication Technology - Year 2014, Vol 3, Issue 1
Abstract
Large numbers of real world applications are producing continuous and large data streams for further analysis. Data stream mining is allowing users to extracts knowledge in the form of patterns or model from the continuous data streams of information. With limitations of data stream mining, it has many challenges in storage, computation. Also the data streams are subjected to changes with environment and time. So while designing the data stream mining process it is important to have flexibility to promptly adapt with changing concepts and data distribution depending on nature of data. The Reinforcement learning is an approach to learning by interacting with the environment. This method builds better tree over time as it learns from the feedback provided by the previous building operations. The objective of this article is to analyze the various classification applications used as one of data mining techniques. Also we reviewed reinforcement learning framework, methods of reinforcement learning, and building a decision tree with reinforcement learning for time changing data.
Authors and Affiliations
Pramod Patil, Ashwini Ahirrao
Keywords
Related Articles
- EP ID EP27822
- DOI -
- Views 236
- Downloads 0
Estimation of Complexity of the Regenerator Placement Problem in Optical Networks
Placement of regenerators in optical networks has attracted the attention of recent research works in optical networks. In this problem, we are given a network with an underlying topology of a graph and with a set o...
Information Mining Thorugh Big Data
Big Data is another term used to recognize the datasets that because of their extensive size and multifaceted nature. Huge Data are currently quickly extending in all science and building spaces, including physical,...
Live Activity Index To Gather Effective Live Traffic Broadcasting To Users
A few online services give live movement information by dissecting gathered information from street sensors, activity cameras, and group sourcing strategies, for example, Google-Map, Navteq, INRIX Traffic Information...
Design and Implementation of Body Movement Recognition System
In this project body movement recognition system based on the Non-vision over the Zigbee Wireless Sensor Network. Now a day's always (day and night time) vision based coma, paralysis patient’s posture information can...
Power Management Strategy In Hybrid PV-FC And Wind- Power Generation Systems By Using Multi Input Single-Control (MISC) Battery
Global energy consumption is dramatically increasing due to higher standard of living and the increasing world population. The world has limited fossil and oil resources. As a consequence, the need for renewable ener...