IOSR Journal of Computer Engineering (IOSR-JCE) e-ISSN: 2278-0661, p- ISSN: 2278-8727Volume 16, Issue 3, Ver. IV (May-Jun. 2014), PP 55-58 www.iosrjournals.org www.iosrjournals.org 55 | Page Efficient Weather Prediction By Back-Propagation Algorithm Manisha Kharola 1 and Dinesh Kumar 2 1 M.Tech Scholar of computer science & Engineering , SRCEM, Palwal 2 Department of computer science & Engineering , SRCEM, Palwal Abstract: Artificial Neural Networks (ANNs) have been applied extensively to both regress and classify weather phenomena. While one of the core strengths of neural networks is rendering accurate predictions with noisy data sets, there is currently not a significant amount of research focusing on whether ANNs are capable of producing accurate predictions of relevant weather variables from small-scale, imperfect datasets. Our paper makes effort to use back propagation algorithm to train the network. So, that it can help in predicting the future weather. Keywords: ANN, Back Propagation, Weather Prediction. I. Introduction Weather process is a dynamic and non linear phenomenon. There is a need to apply statistical post- processing techniques on modeled forecast fields to improve the prediction quality and value. India is an agricultural country so weather prediction plays a very important role. If we knew that what will be the weather in future then we can work according to that. The weather prediction should be accurate and more precise. An Artificial Neural Network (ANN)is an information –processing paradigm that is inspired by the way biological nervous system such as the brain, process information. It is composed of a large number of highly interconnected processing elements(neurons) working in parallel to solve the specific problems. The weight on the connections encodes the knowledge of a network. Each neuron has the local memory and the output of each neuron depends upon only the input signals arriving at the neuron and value in neuron’s memory. The intelligence of a neural network emerges from the collective behaviour of neurons, each of which performs only very limited operations. Even though each individual neuron works slowly, they can still quickly find the solutions by working in parallel. Neural networks, with their remarkable ability to derive meaning from complicated or imprecise data, can be used to extract patterns and detect trends that are too complex to be noticed by either humans or other computer techniques. Using ANN has following benefits:-  No linearity- The answer from the computational neuron can be linear or not.  Adaptive learning- The ANN is capable of determine the relationship between the different examples which are presented to it without requiring a previous model.  Self-organization- This property allows the ANN to distribute the knowledge in the entire network structure, there is no element with specific stored information.  Fault tolerance- This can be shown in two senses: The first is related to the samples shown to the network, in which case it answers correctly even when the examples exhibit variability or noise; the second appears when in any of the elements of the network occurs a failure, which does not affect its functioning due to the way in which it stores information. Back propagation is a method of training multilayer ANNs which use the procedure of supervised learning. Supervised algorithms are error-based learning algorithms which utilize an external reference signal(teacher) and generate an error signal by comparing the reference with the obtained output. Based on error signal, neural network modifies its synaptic connection weights to improve the system performance. In this scheme, it is always assumed that the desired answer is known a priori[2]. The traditional Back Propagation Neural Network(BPNN) Algorithm is generally used in solving many practical problems. The BPNN learns by calculating the errors of the output layer to find the errors in the hidden layers. Back-Propagating is highly appropriate for problems in which no relationship is found between the output and inputs. Due to its suppleness and learning capability, it has been effectively implemented in wide range of applications[3]. II. Literature Survey Many works were done related to the weather forecasting system and BPNN. They are summarized below. Singh, Bhambri and Gill[5] has highlighted on the temperature prediction is a temporal and time series based on process. Due to non linearity in climatic physics, neural network are suitable to predict these meteorological processes. Back Propagation integrated with genetic algorithm is the most important to train the