RAINFALL PREDICTION USING MACHINE

LEARNING
2022 First International Conference on Artificial Intelligence Trends and Pattern Recognition (ICAITPR) | 978-1-6654-2521-6/22/$31.00 ©2022 IEEE | DOI: 10.1109/ICAITPR51569.2022.9844203

Akash Gupta Hitesh Kumar Mall Janarthanan.S

School of Computing Science School of Computing Science School of of Computing
And Engineering And Engineering Science
Galgotias University Galgotias University And Engineering,
Greater Noida, UP, India Greater Noida, UP, India Galgotias University
Email: ag2443593@gmail.com Greater Noida,UP,India

Abstract— Rainfall forecasting is a single ofdifficult and

unpredictable undertakings that has a major influence on human
society. Predictions that are accurate and timely can help to avert
human and financial loss. This study contains a series of
experiments that include the utilisation ofbasic machine learning
techniques to build Weather forecasting models that estimate
whether it will rain in major cities tomorrow based on the day's
meteorological data.This comparative research looks at three
aspects of modelling inputs, modelling methodologies, and pre-
processing procedures.The findings demonstrate how different
machine learning systems perform on a range of assessment
parameters, as well as their capacity to forecast rainfall using
weather data analysis. Agriculture is crucial to India's survival.
The importance of rainfall in agriculture cannot be overstated.
Rainfall forecasting has been a key problem in recent years.
Individuals can be more aware of the weather and make more
educated judgments by predicting rainfall. Fig. 1. Google Trends on Fake News

Index Terms—Rainfall prediction, parameters,
Review analysis, Machine Learning, Prediction System.
I. INTRODUCTION
Rainfall has a crucial part in the evolution of fauna
and plants in the natural world. It is essential for
humans, animals, plants, and other living things.
Water is regarded as one of the world's most valuable
natural resources.and it is utilised extensively in
agriculture and farming. Changing weather
patternsand growing greenhouse gas emissions have
made it more difficult for mankind and the rest of the
globe to get enough rain to satisfy their demands and
continue to use it in their daily lives. As a result, it's
become increasingly important to track shifting
rainfall patterns and forecast rain.Our objective of this
research is to forecast rainfall through arbitrary
forests, vector support classifiers, and other
techniques. Rainfall forecasting may help with not
only the research of shifting rain patterns, but also the
preparation of preventative actions in the event of a
disaster, as well as the crisis management. Global
warming is caused by increased Cfcs emission from
refrigerator, air conditioning units, sprays, and printer,
among other regular items. Temperatures are causing
climate change.
Authorized licensed use limited to: Indian Institute of Technology - Jodhpur. Downloaded on September 25,2023 at 17:38:20 UTC from IEEE Xplore. Restrictions apply.
However, rainfall forecast and weather updates benefit with in handling
of macro-level issues like floods and agricultural challenges caused by
inadequate or severe rainfall, as well as micro-level concerns like
starvation. Rainfall forecasting might also benefit folk's health and
wellbeing by giving them information regarding rainfall patterns and
forecasting rainfall through techniques like spontaneous forest, vector
support classifier, and others. Rainfall predictions can assist individuals
in coping with hot, humid conditions. In the modern world,
technological advancements have opened up more opportunities for
invention and creativity. Despite the fact that the current difficulties are
most probably the outcome of technical breakthroughs, it is vital to
comprehend the variety of possibilities and opportunities that
technology innovation has offered to people.Moreover, inaccurate or
inadequate rainfall forecasting is a cause of worry in water resource
management. Accurate and exact rainfall forecasts may help with not
only the effective and economical use of resources, but also the
preparation of event initiatives and plans. It is vital to develop and
maintain a system which will help in accurate estimates and offer users
with quick access to this information.
II. OBJECTIVE
The objective of study is to implement rainfall prediction with machine
learning techniques such as random forest, Decision tree, logistic
regression , neural network.

III. METHODOLOGY
Response-based prediction often elicits intense feelings, and as
a result, it spreads swiftly on social media. The news'
credibility is determined using a response-based method that
considers past data replies. There are two sections to this
project. The greater posterior probability approach was used to
forecast the weather in the first phase. Regardless of the fact
the accuracy scores improved, we determined that this strategy
was useless.By assessing the behavior to certain weather
inside this second phase, we devised a strategy for more
correctly identifying showers." Several Sci- kit packages were
utilised in this Python-based project. Python has a significant
number of Machine Learning extensions and modules. The
Sci-Kit Learn module is the ideal place to go for ML
evaluation because it offers access to almost all types of
machine learning for Python.
Fig. 2. Methodology
Authorized licensed use limited to: Indian Institute of Technology - Jodhpur. Downloaded on September 25,2023 at 17:38:20 UTC from IEEE Xplore. Restrictions apply.
A. Data Retrieval
Kaggle/Twitter appears to be our main data source. After
you've created a Twitter Developer account,we were granted
privileges to begin harvesting tweets. We used a variety of
keywords to extract tweets on rain, weather, and climate.
Information on the user and the tweet was also retrieved, as
well as the user's ID, username, tweet text, andif the user's
status was checked Our decision variable was the user-verified
column, with the assumption that if a user-verified account
produced a tweet, it was most likely true news. Consumers are
frequently supposed to trust everything communicated
through these verified accounts. As a result, it's natural that
before granting a verified blue tick to a user's profile, Twitter
goes through a rigorous vetting procedure. Users who wish to
validate his accounts must go through a lengthy process in
which Twitter investigates the person's past and authenticates
them. In essence, we collect information from a variety of
sources and edit it ourselves.. Name, text, and label are
included in the provided data set. The header appears in the
title, the text provides the data, and the label indicates if the
information is correct or incorrect.
B. Data Pre-processing
In this study, we're investigating weather forecasting. The
main objective of the research is to apply machine learning to
predict rainfall with high accuracy and precision. As a
consequence, the research contains and analyses the source
data during the pre-processing phase, and then uses this data
during the stimulating or processing phase to forecast the
output in an effective and efficient manner. For statistics,
Erbil, Nicosia, and Famagusta's metrological departments
were consulted. The information from the various
meteorological departments is collected once a year, as this
study will also look at earlier rainfall predictions.Important
rainfall elements are included in the data. Temperatures, wind
direction, prevailing winds, barometric pressure, and moisture
are all taken into account. For Erbil, Cyprus, and Famagusta,
each precipitation parameter is investigated separately. The
statistical data gathered in this study will be used to translate
the result into a rainfall forecast. Machine-learning algorithms
are used to process the data.
C. TDFVectorizer
The acronym TF-IDF stands for Term Frequency Rate. This
is a common method for transforming text into a meaningful
numerical representation, which is then used to train a
machine learning system.
D. Machine Learning Algorithms
Inside this model, we apply a variety of algorithms that are
extremely effective for identifying rainfall forecast and
supporting us in reaching the correct classification result for
the data in the provided data set.We use routinessuch as
random forest, regression models, decision tree, and neural
network.
(i) Random forest algorithm: -Instead of focusing on a
single tree structure, the confusion matrix utilizes this
same forecast by each tree and anticipates the correct
outcome predicated just on majority votes of
predictions." Rf is a classifier that "uses a decision
tree on different sets of a set of data and averaged
nearly them to improve the forecasting ability of that
dataset."'

(ii) Logistic Regression: This technique uses a set of

independent variables to forecast a categorical
dependent variable. Logistic regression is used to
Fig .3. Data successfully show
predict the outcomes of categorical variables. As a
result, the output has to be categorical or discrete in
both cases. It can also be Yes or No, Zero or One,
True or False, and so on, but instead of exact values
like 0 or 1, it gives probability numbers that are
somewhere in the middle.

(iii) Decision Tree algorithm: -Decision Trees are

although supervised learning techniques can be used
to solve classification and regression problems, they
are most typically utilised to solve classification
problems. In this tree-structured classifier, internal
nodes hold dataset properties, branches offer rule
base, and each leaf node represents a conclusion. A
Decision tree is made up of two nodes: the Leaf Node
and the Decision Node..

(iv) Neural Network :- Artificial Neural Network (ANN)

is a computing method that captures a wide range of Fig.4. RainTomorrow indicator (unbalanceDataset).
products that are coupled in some way to allow
connection between the units. It is also known as
"artificial neural network.".

E. Training TestingModel
Following the introduction of the parameters of each
methodnetwork, the data must be tested using supervised
learning. The algorithm use trained data to extract
characteristics from the dataset. The quality of our generated
model is then assessed using performance measures such as
accuracy, reliability, f1 measure, and recollection.It's also
known as the confusion matrix.’

IV. RESULT
The findings of on top on model demonstrate how effective the
response is at distinguishing between true and fake
information. The outcomes of the research of the model's Fig.5. RainTomorrow indicator (balanceDataset).
operation are listed below.

Authorized licensed use limited to: Indian Institute of Technology - Jodhpur. Downloaded on September 25,2023 at 17:38:20 UTC from IEEE Xplore. Restrictions apply.
 VI. CONCLUSION

Rainfall is among nature's most importantoccurrences that has an

impact on all living creatures, not just humans. Because of
shiftingChanges in climate conditions and rainfall cycles are
occurring, as well as the earth's temperature rising.However, a
system that predicts rainfall using artificial intelligence and neural
networks could help solve these problems and meet demand, which
is already common in modern technology, is necessary. The study's
purpose was to create a machine-learning-based prediction method
that could correctly and effectively estimate monthly rainfall with
the least amount of error.The research looked at rainfall data from a
variety of locations as well as different Machine Learning methods
including randomized forest and support vector. To evaluate and
Fig.6, Logistic Regression Accuarcy verify the training data, a comparison of data and expected data is
employed. To obtain a more accurate and precise outcome
prediction, the system employed feature extraction.

VII. ACKNOWLEDGMENT
We required the assistance and direction of a few notable individuals
in order to fulfil our mission, and we owe them our appreciation.
We're quite proud of ourselves for doing our task. Mr..S Janarthanan,
Project Mentor, University of Galgotia's, is to be thanked for exposing
us to work methodology and providing us with excellent assignment
aid. Throughout many consultations, his regard for the "surrounding
structures" was evident.We'd want to thank everyone who has assisted
us, either directly and indirectly, in completing this project. Many
individuals, especially students and members of the team, have
offered constructive feedback on this proposal, motivating us to better
our objective. We appreciate everyone's direct and indirect
Fig.7. Decision Tree Accuarcy contributions to the completion of our project. Finally, we'd want to
express our gratitude to our kind parents for everything they've
We got different accuracy by using different algorithms, as accomplished for us.
you see within the outcome.
REFERENCES

1. Abhishek,K.,Kumar,A.,Ranjan,R.,&Kumar,
V. ANALYSIS THERESULT
S. (2012). A Rainfall Prediction Model using
Artificial Neural Network. IEEE Control and
The graph here depicts the model's accuracy in determining System Graduate Research Colloquium, 1-5.
if it is perfect. The model is absolutely hooked to accuracy in
machine learning, meaning that our model is perfectly good if 2. Agnihotri, G., & Panda, J. (2014). Comparison
accuracy is nice. Using these algorithms, we were able to get a of Rainfall from Ordinary and Automatic Rain
fairly accurate rainfall prediction. We ran these algorithms Gauges in Karnataka. Mausam, 65(4),1-8.
several times to observe how the accuracy varied, but the
results have always been accurate. 3. MuttalebAlhashimi, S. A. (2014). Prediction ofMonthly
Rainfall In Kirkuk Using Artificial Neural Network and
Time Series Model. Journal of Engineering and
Development, 18(1),1-12.

4. Mittal, P., Chowdhury, S., Roy, S., Bhatia, N., &

Srivastav, R. (2012). Dual Artificial Neural Network for
Rainfall-Runoff Forecasting. Journal of Water Resource
and Protection, 4(12),1-4.

Table 1. Accracy Table

Authorized licensed use limited to: Indian Institute of Technology - Jodhpur. Downloaded on September 25,2023 at 17:38:20 UTC from IEEE Xplore. Restrictions apply.
 5. Deepak Ranjan Nayak, Amitav Mahapatra and
Pranati Mishra, "A Survey on Rainfall Prediction
using Artificial Neural Network", International
Journal of Computer Applications, vol. 72, no. 16,
pp. 0975-8887, June2013

6. Nicholls, N.: Atmospheric and climatic hazards:

Improved monitoring and predictionfor disaster
mitigation. Natural Hazards 23(23), 137155(2001)

7. Omotosho, J. B., Balogun, A., &Ogunjobi, K.

(2000). Predicting monthly and seasonal rainfall,
onset and cessation of the rainy season in west
africa using only surface data. Internationaljournal
of climatology, 20,1-16

8. Shaikh, L., &Sawlani, K. (2017). A Rainfall

Prediction Model Using Articial Neural Network.
International Journal of Technical Research and
Applications, 5(2),1-3.

9. Sivakumar, M. V. (2006). Climate prediction and

agriculture: current status and future challenges.
Climate research, 33(3-17),1-12.

10. Qian, W., Kang, H.-S., & Lee, D.-K. (2002).

Distribution of seasonal rainfall in the East Asian
monsoon region. Theoretical and Applied
Climatology, 73(3-4), 1-9.

Authorized licensed use limited to: Indian Institute of Technology - Jodhpur. Downloaded on September 25,2023 at 17:38:20 UTC from IEEE Xplore. Restrictions apply.