Special Issue - 2017 International Journal of Engineering Research & Technology (IJERT)

ISSN: 2278-0181
ICIATE - 2017 Conference Proceedings

Weather Forecasting using Neural Network

Priyanka Mahajan∗, Chhaya Nawale†, Siddheshwar Kini‡, Prof. Krishnanjali Shinde§
∗†‡§ Department of Computer Engineering, Atharva College of Engineering
University of Mumbai, Mumbai-400095, India

Abstract - Weather Forecasting determines future state of acceptable models for weather prediction. Also, computing
the atmosphere. Accurate weather forecasting is very for these prediction models is very expensive because of
important because agricultural and industrial sector are based compute intensive nature.
on it. We are presenting weather predictions using Artificial On the contrary, data mining models works on historical data,
Neural Network and Back Propagation Algorithm. We are it works on probability and/or similarity patterns. For all the
implementing data intensive model using data mining prediction categories, the model works in similar fashion, and
technique. Weather is a dynamic and non-linear process and
expects to return the moderate accuracy [9].In weather
artificial neural network (ANN) can deal with such type of
Process. The research also states that ANN is the best approach forecasting output of the model may be required for daily
than traditional and numerical methods. Data mining technique weather guide or weekly or monthly weather planning. Thus,
with neural network gives useful information for weather the accuracy of the result is a very important aspect in
prediction which reduces cost as compare to other prediction forecasting so as to give the best result among all other
models. The focus is mainly on neural network with back weather forecasting model.
propagation technique for daily weather forecasting.
II. LITERATURE REVIEW
Keywords: Neural Network, Back propagation Algorithm,
Daily Weather Forecasting, ANN, Weather Prediction. Paras, Sanjay Mathur, Avinash Kumar, and Mahesh Chandra
I. INTRODUCTION (2009) they eliminate the need of the satellite based systems
for weather forecasting which is costlier than data mining
Weather forecasting is a process of identifying and system and require complete support system. This model used
predicting to certain accuracy the climatic conditions using ANN with time series features [4].
multiple technologies. Many of the live systems rely on
weather conditions for making necessary adjustments in the Mohsen Hayati [5] states that ANN for one day ahead
systems. Quantitative forecast like temperature, humidity prediction of temperature. Here multilayer perceptron (MLP)
and rainfall are important in agriculture area, as well as to is trained for ten years meteorological data from Iran which
traders within commodity markets [6]. Nowadays there are was divided into four seasons namely spring, summer, fall
many approaches we are using for weather forecasting. and winter. MLP is a network of three layers having sigmoid
Mathematical modeling, statistical modeling and artificial transfer function for hidden layers and linear transfer function
intelligence techniques are some of them. Mathematical for output layer. Hidden neurons were decided using the trial
models of the atmosphere to predict future weather based on and error method. It was concluded that MLP has minimum
current weather conditions is numerical weather Prediction. prediction error, good performance and reasonable prediction
accuracy. They have used historical data for ten years and
It needs full knowledge of atmospheric dynamics and achieve minimum error and reasonable accuracy.
involves calculations with a large number of variables and
datasets. Due to the advancement of modern computer Ch.Jyosthna Devi,B.Syam Prasad Reddy, K. Vagdhan
hardware there have been many improvements in numerical Kumar, B. Musala Reddy, N.Raja Nayak this gives
weather prediction [1]. We are using ANN which is based on Information about Artificial neural network using Back
smart analyzing the trend from historical data. The other Propagation Neural Network (BPN). This method was more
models are accurate in calculation but not in predictions the efficient than numerical differentiation [7].
can not able to adapt the irregular patterns of data which can
neither be written in form of function or deducted as formula Gyanesh Shrivastava, Sanjeev Karmakar, Manoj Kumar
[2]. Kowar, Pulak Guhathakurta given that BPN, RBFN is
efficient model to forecast monsoon rainfall as well as other
The weather forecast reports needs some intelligent weather parameter prediction over the smaller geographical
computing which can able to read the nonlinear data and region.BPN and RBFN were given appropriate solutions for
generate some rules and patterns to study from the observed prediction of long-range weather forecasting [8].
Data to predict the weather in future. Use of ANN will give
more accurate result. Here, the error may or may not reduce Arti R. Naik, Prof. S.K.Pathan [11] has given the limitation
completely. But, the accuracy will improve as compared to about some existing weather forecast model and benefits of
previous forecasts. Weather research and forecasting (WRF) artificial neural network. The focus is on different methods of
model, General Forecasting Model, Seasonal Climate weather prediction using artificial neural networks and
Forecasting, Global Data Forecasting Model, are currently training of this network using Backpropagation algorithm.

Volume 5, Issue 01 Published by, www.ijert.org 1

Special Issue - 2017 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ICIATE - 2017 Conference Proceedings

this, we are investigating the use of data mining techniques in

R Lee [10] proposed an innovative, intelligent, multi-agent forecasting maximum temperature, rainfall, evaporation and
based environment called as intelligent Java Agent wind speed. This was carried out using Artificial Neural
Development Environment (iJADE). It is used for weather Network and Back propagation Algorithm and
prediction of eleven weather stations in Hong Kong using meteorological data collected from previous 3 years.
five years data which provides more than 7000 data records.
The model uses GA for input node selection, a fuzzy Data Collection: the data used for this work was collected
classification for rainfall parameters and neural network for from different official websites and Metrological Department.
training using a BPN. Its experimental results are more The case data covered the period of 36 months. The following
accurate than single point sources using similar network and procedures were adopted at this stage of the research: Data
other networks like Radial Basis Function Network, Cleaning, Data Selection, Data Transformation and Data
Learning Vector Quantization and Naive Bayesian Network. Mining.
Data Transformation: this step is also known as data
III. ARTIFICIAL NEURAL NETWORK FOR consolidation. It is the stage in which the selected data is
WEATHER FORECASTING transformed into forms appropriate for data mining. The data
file is saved in Commas Separated Value (CVS) of file
In any Forecasting accuracy is very important .The input format and the datasets were normalized to reduce the effect
parameters for a weather forecasting model are different of scaling.
types of data required different types of methods; and need Data Mining Stage: It is divided into three phases. At each
to be handled accordingly. Statistical methods are associated phase the algorithm is used to analyze the meteorological
with datasets. The testing method adopted for this research was
Linear data on the other hand Artificial Intelligence methods percentage split that train on a percentage of the dataset, cross
are associated with nonlinear data [12]. Different types of validation on it and test on the remaining percentage.
learning models of Artificial Intelligence are genetic Thereafter interesting patterns representing knowledge were
algorithms, neuro-fuzzy logic and neural networks. Among identified.
which neural networks is more preferred.
Weather forecasting can be done more accurately using On the other side on user device we are creating the Android
ANN. Because daily weather data has multiple parameters Application where user need to register him so that daily
such as temperature, humidity, rainfall amount, cloud predicted weather notification will be send on his device.
distance and size, wind speed and direction, etc. All these
parameters are non linear, but they required to be processed
together to determine temperature, rainfall, humidity or
weather status for the future day. Such type of applications
need complex models and can able to produce the required
result by generating the patterns on its own by performing
self-learning using the training data given to the model.
To develop an ANN model for weather forecasting, region
selection for input data and parameters is necessary. The
input data is to be taken from a specific area on which the
model is trained and tested so that the model is able to
generate accurate results. The number of input data given to
model also helps to improve accuracy of the model by giving
the results with a high degree of similarity between predicted
and actual output data. The available data may be noisy thus,
data should be cleaned. Similarly, it has to be normalized
because, all the parameters are of different units and
Figure 1: Methodology
normalization will help the input and output parameters to
correlate with each other [6]. The data should be divided in
A. System Phase:
training and testing samples in proper proportion so that the
results can be predicted, tested and validated properly. This system uses the concept of Weather Prediction. The
Structure of the NN model also has a great impact on Three phases are as follows:
generation of accurate results. The multilayer ANN helps in
predicting nonlinear data more efficiently. The activation B. Registration Phase:
function will be different for different layers of NN as per
need. In registration process, the user will require to fill all the
details such as name, user id, email id, phone number etc. He
IV. METHODOLOGY
also requires registering his location so as to get the
notification according to location.
Weather forecasting is an application in meteorology and has
been one of the most scientifically and technologically
challenging problems around the world in the last century. In

Volume 5, Issue 01 Published by, www.ijert.org 2

Special Issue - 2017 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ICIATE - 2017 Conference Proceedings

C. Login Phase:

In login phase, the admin login to the system which is not

allowed to user. After logging in the system the
preprocessing of data is done by admin.
D. Notification Phase:

User will get daily weather notification using application

which can be installed by the user.

V. BACK-PROPAGATION APPROACH

The back propagation algorithm is used in layered feed-

forward ANNs. It uses supervised learning, which means the
model trains itself with the use of target output. For every set
of input data the target output is provided. The neural
network model processes the input data with random values
for weights and suitable activation function using one or
more hidden layer in between and then produces the
predicted output. This predicted output is then compared
with the target output provided for same input dataset. Thus,
error is calculated by subtracting predicted output from
target output. Using this error, the weights are adjusted and
again the entire process is repeated for multiple epochs until
the error is minimal or in acceptable range [3].
Figure 2: Back Propagation Approach
We start the training with random weights, and the goal is to
adjust them so that the error will be minimal. The area for Seventy percentages of the dataset will be used for training
input data can be any one of a meteorological station area in and the other thirty percentages of the dataset will be used for
which all the data is limited to a certain region. The different testing and validation. Hidden layers are required for
input parameters are taken viz. temperature, relative processing nonlinear data. Better results can be achieved with
humidity, air pressure, wind speed and direction, cloud high accuracy when learning rate is smaller but its
amount and height, rainfall, etc. performance is slower. Activation functions are applied on
each neuron to get the output of neuron on a given input in
Input data is then pre-processed and cleaned. That means it the neural network. The sigmoid function is a special case of
is checked with any outlier and that is removed, missing logistic function which has a sigmoid curve. The sigmoid
values are entered, and data is checked if it is in the given transfer function can be used for hidden layers and for the
range for the given parameter. Later ANN is designed with output layer the linear transfer function can be used.
number of input and output nodes, hidden layers, activation The correctness and accuracy of the model can be checked
function, and maximum number of epochs, weights, bias, using the Mean Squared error (MSE) function. The MSE
goal and learning function. Neural network is trained with measures the average of the squares of errors that is, the
seventy percentages of the input data. Where the model is difference between the actual output and the predicted output
trained using this observed data to forecast the weather, of the model. Lesser the MSE value of the model, more
followed by testing done using remaining thirty percentages accurate the results are.
of input data. Then the mean squared error and accuracy is
calculated for the model by comparing the output of testing
with target output. This model generates output in terms of VI. FUTURE SCOPE
minimum and maximum temperature of the day, relative
humidity and rainfall. The accuracy from weather forecasting model using ANN
and Back Propagation Algorithm is more than other
Statistical Model. An extension to this technology can be
done using any of the other technique instead of Data mining
and Different Algorithm.

Volume 5, Issue 01 Published by, www.ijert.org 3

Special Issue - 2017 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ICIATE - 2017 Conference Proceedings

VII. CONCLUSION [5] Mohsen Hayati and Zahra Mohebi, “Temperature Forecasting based
on Neural Network Approach”, World Applied Sciences Journal
2(6): 613-620, 2007, ISSN 1818-4952, IDOSI Publications, 2007.
The different methods for weather forecasting are reviewed.
[6] R Lee, J Liu, "iJADE Weather MAN: A Weather Forecasting System
ANN with back propogation is recommended for weather Using Intelligent Multiagent-Based Fuzzy Neuro Network", IEEE 181
Forecasting. ANN with back propagation uses an iterative Transactions on Systems, Man and Cybernetics - Part C: Applications
process of training where, it repeatedly compares the and Reviews, vol 34, no 3, 369 - 377, August 2004.
observed output with targeted output and calculates the error. [7] Ch.Jyosthna Devi #1, B.Syam Prasad Reddy#2, K.Vagdhan
This error is used to readjust the values of weights and bias Kumar#3,B.Musala Reddy#4,N.Raja Nayak#5, 2012, ANN Approach
for Weather Prediction using Back Propagation, International Journal
to get an even better output. Hence this method tries to of Engineering Trends and Technology-Volume3Issue1- 2012, ISSN:
minimize the error. Thus, Artificial Neural network with 2231-5381.
Back propagation algorithm seems to be most appropriate [8] Gyanesh Shrivastava, Sanjeev Karmakar, Manoj Kumar Kowar, Pulak
method for forecasting weather accurately. Guhathakurta, 2012, “Application of artificial neural networks in
weather forecasting: A comprehensive literature Review”,
International Journal of Computer Applications (0975 – 8887),
The weather Forecasting has a big challenge of predicting Volume 51– No.18.
the accurate results which are used in many real time [9] Folorunsho Olaiya, Owo, “Application of Data Mining Techniques in
systems like electricity departments, airports, tourism Weather Prediction and Climate Change Studies”, IJIEEB Vol.4,
centers, etc. The difficulty of this forecasting is the complex No.1, February 2012.
nature of parameters. Each parameter has a different set of [10] Arti R. Naik*, Prof. S.K.Pathan**Weather classification and
ranges of values. This issue is addressed by ANN. It accepts forecasting using feed forward neural network, International Journal
of Scientific and Research Publications, Volume 2, Issue 12,
all complex parameters as input and generates the intelligent December 2012 1 ISSN 2250-3153.
patterns while training and it uses the same patterns to [11] Gill J. “Training back propagation neural networks with genetic
generate the forecasts. Algorithm for weather forecasting” IEEE 8th international symposium
on Intelligent systems and informatics Serbia, 2010.
[12] R.Sallehuddin,"Forecasting Time Series data using Hybrid Grey
REFERENCES Relational Artificial Neural Network and Auto Regressive Integrated
Moving Average," Journal of Applied Artificial Intelligence, vol. 23.
[1] P. Lynch, (2008) “The Origins of Computer Weather Prediction and
Climatic Modeling", Journal based on Computational Physics, vol.
227, 3431-3444
[2] Kumar Abhishek, Abhay Kumar “A Rainfall Prediction Model Using
Artificial Neural Network” department of Information Technology
IEEE Control and System Graduate Research Colloquium (ICSGRC
2012), pp 82-87.
[3] Carlos Gershenson,“Artificial Neural Network for Beginners”,
[email protected], university of Sussex.
[4] Paras, Sanjay Mathur, Avinash Kumar, and Mahesh Chandra, 2009,
A Feature Based Neural Network Model for Weather Forecasting
International, Journal of Computational Intelligence Volume 4
Number3.

Volume 5, Issue 01 Published by, www.ijert.org 4