International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 03 Issue: 06 | June-2016 www.irjet.net p-ISSN: 2395-0072 © 2016, IRJET | Impact Factor value: 4.45 | ISO 9001:2008 Certified Journal | Page 702 Forest Fire Detection Through Various Machine Learning Techniques using Mobile Agent in WSN Anupam Mittal 1 , Geetika Sharma 2 , Ruchi Aggarwal 3 1 Anupam Mittal, Dept. Of Compuetr Science & Engineering, Chandigarh University, Mohali,India 2 Geetika Sharma, Dept. Of Compuetr Science & Engineering, Chandigarh University, Mohali,India 3 Ruchi Aggrawal, Dept. Of Compuetr Science & Engineering, Chandigarh University, Mohali,India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract - Wireless sensor networks monitor dynamic environments that change suddenly over time. Machine learning also inspires many practical solutions that less energy consumption and to increase network lifetime. This paper provides review of machine learning techniques for detection of forest fire in wireless sensor network. Key Words: Wireless Sensor Network, Machine Learning Techniques, SVM, ANN, DT, FFNN. 1.INTRODUCTION Forests play an important role for supporting the human environment and Forest fires are among the largest dangers for forest preservation. Wireless Sensor Networks are used to forest fire detection. Wireless sensor networks have been used in a variety of applications such as:  Habitat monitoring  Forest fire detection  Event detection  Health and medical monitoring  Target tracking  Surveillance monitoring Event detection is one of the applications of data observation in wireless sensor networks. A large amount of sensor data is semantically processed and only relevant information is sent to the user [14]. 2.RELATED WORK Aditi, Yashwant and V. Mohindru [1] provide new approach how regression works best for forest fire detection with high accuracy by dividing the forest fire dataset. They also present Comparsion of various machine learning techniques like as SVM (support vector machines), neural network, decision tree, regression, so on for detection of forest fires and new approach perform better as compared to other machine learning techniques. Archana and Dr.Upadhyay [2] introduce a force based algorithm to auto deploy a sensor network to low consumption of energy by sensor nodes and to increase their network lifetime. M.Oladimeji, M.Turkey and S.Dudley [3] said that hybrid approach of k-means clustering. They are uses three classification approaches like as the FFNN, Naïve Bayes and decision tree make hybrid approach for detection of fire with high accuracy. K .Trivedi and A. Srivastava provide a framework that centralizes the use of mobile agent in wireless sensor network that can help in quickly detection of forest fire and monitoring of it with less energy consumption [4]. K. Kim et.al. Proposed new feature selection methods random forest-forward selection ranking (RF-FSR) and random forest-backward elimination ranking (RF-BER) [5]. Y.Singh, S.Saha and U.Chugh, C.Gupta [6] proposed an ensemble distributed machine learning approach for event detection and it perform in two steps, base step and Meta step. They also used clustering and SVM approaches for detection and prediction of event.