A Novel System for Recommending Agricultural Crops Using Machine Learning Approach

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 11 Issue: 02 | Feb 2024 www.irjet.net p-ISSN: 2395-0072
© 2024, IRJET | Impact Factor value: 8.226 | ISO 9001:2008 Certified Journal | Page 85
A Novel System for Recommending Agricultural Crops Using Machine
Learning Approach
Prasad Raghunath Mutkule1
1Assistant Professor, Department of Information Technology, Sanjivani College of Engineering, Kopargaon,
Maharashtra
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Farmers play a crucial role in our lives. It takes
usually a couple of months in advance for a crop
recommendation system to predict crop yields for improved
production. Crop prediction is dependent on the use of
computer programs that describe in quantifiable terms the
interactions between plants and their environment, as well as
their soil characteristics. It is first necessary to collect a soil
sample from the field for soil testing. The Maharashtra state's
agricultural production is decreased as a result ofuncertainty
caused by its coastal location. Productivity should increase as
population and area grow, but it cannot. Asaresultofclimatic
factors, farmers are no longer able to use word-of-mouth.
Farmers can benefit from good agricultural information
thanks to the growth of IT in the world. Therefore, in this
present climate, there is a need for intelligent use of modern
technology in the agribusiness sector. Crop prediction is a
challenging task for farming because it depends on feature
selection and classification. Using a novel framework is
proposed for choosing crop characteristics, assessing crop
yields through classification. The categorization algorithms
employed in this approach differ from those used in other
research, as they employ multiple prediction methods.
Key Words: Agriculture, Crop Prediction, Classification,
Machine Learning, Segmentation
1.INTRODUCTION
There are few countries in the world that still practice
agriculture as old as India. As a result of globalization,
agriculture trends have changed dramatically in recent
years. In India, agriculture has been adversely affected by a
variety of factors. Health has been regained through the
development of many new technologies.Theuseofprecision
agriculture is one such technique. India is experiencing a
boom in precision agriculture.
"Site-specific" farming is precision agriculture. The
technology has helped to achieve efficient inputs, outputs,
and better decisions as far as farming is concerned. There
are many systems availablethatcandeterminetheinputsfor
a specific plot of land. Precision agriculture has brought
about some improvements, but it has still faced some issues.
In addition to crops and fertilizers, systems can also suggest
farming techniques. Precision agriculture involves the
recommendation ofcrops.Variousparametersaretakeninto
account when recommending crops.
The economy of a country is dramatically impacted by
agriculture. Natural factors are causing Agriculture farming
to degrade in modern times. It is directly dependent on the
environmental factors such as sunlight, humidity, soil type,
rainfall, maximum and minimum temperatures, crop
fertilizers, pesticides, etc. To flourish in Agriculture, one
need knowledge of proper harvesting. India has seasons:
1. From December to March, there is winter
2. From April to June, there is Summer
3. Between July and September is monsoon season
4. During October and November, the post-monsoon
season occurs.
Various seasons and rainfall make it necessary to determine
which crops are suitable for cultivation. The management of
crops, the production of crops, and the productivityfromthe
crops pose major challenges to farmers. As more and more
young people are interested in agriculture these days,
farmers and cultivators need proper assistance regarding
crop cultivation. Assessment of real-world problems by IT
sector is increasing at a faster rate. The amount of data in
agriculture is increasing every day. Agricultural data can be
accessed with the advancement of the Internet of Things
(IoT). In order to extract or use useful information from the
spreading data on agriculture, there is a need for a system
which analyzes agricultural data in an obvious way.
By combining datawithmachinelearningtechniques,wecan
develop a model that can be predictedbasedonthedata.The
solution to farming issues such as crop prediction, rotation,
water requirements, fertilizer requirements, and protection
can be found here. A reliable method forcropcultivationand
management is necessary since there are variable climatic
factors in the environment [14]. Agriculturalists canusethis
information to improve their farming. By using data mining,
farmers can receive recommendations to grow their crops.
Based on climatic and quantity factors, such an approach is
implemented. Agricultural databases can be analysed with
Data Analytics. Based on productivity and season, the crop
dataset has been analysed.

2. MOTIVATION
Farming plays an important role in everyday life.
Classification and feature selection play a crucial role in crop
prediction in farming. The study predicts crop yield by
selecting characteristics from a crop, which is thenclassified.
Our study uses anumberofcategorizationalgorithmsinstead
of just one prediction approach as has been used in other
studies.
3. LITERATURE REVIEW
Shreya S. Bhanose, Kalyani A. Bogawar et al.[1] stated as ,
Farmers require support with their decisionsto enhancethe
quality of their farming by utilizing well-defined and
systematic approaches to predict crops and yields. Since
crop knowledge-bases are not readily available, predicting
the best crops is complex. Better quality farming and higher
revenues can be achieved through crop prediction. To
extract useful information and give predictionsinthefield of
data mining, it is efficient to use data clustering.Inthepast,a
number of approaches have been used for crop prediction.
Making correct decisions with crop prediction model.
Indeed, this helps farmers generate better revenue and
improve farming quality. As a result of a random selectionof
an initial cluster center and a decision regardingthe number
of clusters, traditional clustering algorithms such as k-
means, improved rough k-means make the task ofclustering
more complex. Due to the initial cluster-centric selection,the
modified K-Means algorithm improves accuracyofa system.
Tripathy A.K et al.[2] elaborated,Indevelopingcountrieslike
India, agriculture is one of the most important applications.
Decisions related to agriculture are often based on data
mining. Data mining is the process of extracting relevant
knowledge from a set of data and convertingitintoa human-
understandable format. Climate can have a huge impact on
crop productivity in certain agriculture regions. Crop
management depends on climatic conditions. In order to
achieve good crop management, real-time weather data can
be useful. In order to obtain knowledge and trends,
information and communications technology can be used to
automate the extraction of significant data,which eliminates
manual processes and makes it simpler to extract data
directly from electronic sources. Producing less costs,
increasing yields, and raising market prices can be achieved
in this way. Data mining is also used in analyzing and
predicting useful patterns from vast yet dynamically
changing climatic data. Researchers and engineers have
developed fuzzy logic, artificial neural networks, genetic
algorithms, decision trees, and support vector machines for
studying soil, climate conditions, and water regimes that
affect crop growth and pest control. There is a summary as
well of data mining techniques, neural networks, support
vector machines, big data analysis, and soft computing in
agriculture in this paper.
Ramesh Babu Palepu at al.[3] said that, In developing
countries such as India, agriculture is a backbone for
fulfilling global food demands. It is possible to improve
cultivation yields by applying data mining techniques to
agriculture, especially soils, by revising pledge making
situations and improving pledges. Several issues related to
agriculture require soil analysis for resolution. There are
several data mining techniques discussed, along with their
related work, by several authors in context to soil analysisin
this paper. Soil analysis uses very current data mining
techniques.
Rajeswari et al.[4] given as, Agriculture reliesheavilyonsoil.
Using data mining classification techniques, the work will
predict soil type. Methods used are JRip, J48, and Naive
Bayes are used to predict soil type through data mining
classification techniques. Two types of soil are taken into
consideration whenusingtheseclassifieralgorithms,namely
Red and Black soils. This data can be better modelled using
JRip and Kappa Statistics were raised in the forecast.
Vikas Kumar et al.[8] given in their paper, CT has become a
primary need for humans due to the evolution of Web 2.0.
Farmers lack agricultural knowledge. Farmers and experts
can communicate through ICT. Using spatial data and
agricultural knowledge bases, a semantic web architecture
for generating agricultural recommendationsisproposed by
the authors. In response to climate conditions and
geographic data, our knowledge base will send
recommendations to farmers. In order to find out
information regarding a specific crop, farmers send queries
to a query engine. GIS data and crop knowledge bases may
be accessed in a query. On a mobile device, the query's
results are displayed.
Additionally, it was found that data mining aids in the
analysis and prediction of useful patterns from vast and
constantly changing climatic data. Fuzzy logic, artificial
neural networks, genetic algorithms, decision trees, and
support vector machines have been developed by
agricultural and biological engineers to study soil, climate,
and water conditions related to crop growth and pest
management. This study summarizes the use of Soft
Computing, Big Data analyses, Neural NetworksandSupport
Vector Machines in the agriculture field based on weather
conditions. [10].
Table -1: Review of Existing Methods
Author
and Year
Methodology Advantages Limitations
Shreya S.
Bhanose
et al. [1]
2021
Modified K-
Means algorithm
Improves
Accuracy
Usinglimited
training
data, for
training
Tripathy
A.K et
ANN with Fuzzy
system
Eliminates
manual
Limiting
parameters

Author
and Year
Methodology Advantages Limitations
al.[2]
2019
extraction of
the data
and their
effects
Ramesh
Babu
Palepu at
al.[3]
2019
Data Mining
Helps to
Improve
cultivation
yields
Sample Size
Rajeswari
et al.[4]
2018
JRip, J48 and
Naïve Bayes
Easy
classification
will be done
Further
improvemen
t and
reduction of
computation
time
A.
Swarupa
Rani et al.
[5]
2020
Enhanced SVM
Elaborates
the
Applicationin
terms of
agricultural
field
Memory
limitations
Vikas
Kumar et
al.[8]
2021
IoT Based Web
2.0
Uses spatial
data and
agricultural
knowledge
bases
Not
Applicable
RameshA.
Medar et
al.[11]
2019
Data Mining
Techniques
Detailed
Examination
of the Various
Techniques
are carried
out
More
Computatio
nal time
needed
4. EXISTING METHODOLOGY
Using the data mining process, Tripathy et al. describedhow
pesticides can be managed during crop cultivation. In order
to apply a spatiotemporal analysistocropestimation,Pritam
Bose developed an SNN model. Modified k-means clustering
algorithm was used by Deshpande Radhika and others [13]
to predict harvest and water requirements for crops.
4. 1 Drawback of Existing Approach
 There was no consideration ofother parameters or states
in the existing system.
 Build time is slower.
 Complicated.
 Costly from a computational point of view
5. PROPOSED METHODOLOGY
A number of agricultural parameters influence crop
production. A recommendation can be made to the farmer
based on previous years' cropproduction. Thefarmerwill be
able to determine whether a certain crop has beenyieldinga
good yield recently by this kindofsuggestion.Several factors
can reduce crop production, including crop disease, water
problems, and many others. Farmers may be able to obtain
useful information about the crop in high demand on the
market during that year while examining the production.
The farmer can use this information to determine the trend
in crops in recent years. Based on the crop production
season, farmers will receive recommendations.
5. 1 Advantages of Proposed Approach
 As opposed to the existingsystem,whichconsidersa
single state, our proposed system considers all the
states of India.
 It is possible to extract these recommendations for
educating the farmers. Farmers are given a better
understanding of cropstocultivatethroughpictorial
representations.
 Normalization and scaling are not required
 Interpretation is simple and easy to understand
Fig -1:Proposed Method
5. 2 Machine Learning
Machine Learning is the process of learning from examples
without being explicitly programmable. The purpose of
machine learning is to predict an outcome that can be used
to make useful decisions by combining data with statistical
tools. Data-driven machines can produce accurate results by
learning from data (i.e., example). A close relationshipexists
between machine learning and data mining [14]. Using an

algorithm, the machine formulates answers from data.
Recommendations are typically provided by machine
learning. In order to personalize recommendations, tech
companies use unsupervised learning.
Fig -2: Traditional Modeling vs Machine Learning
(Source:https://www.azavea.com/blog/2017/09/21/buil
ding-inspection-prediction/)
5.3 Working of ML
All the learning happens in machinelearning.Machineslearn
similarly to humans. Experienceisthekeytohumanlearning.
We are better able to predict the futurewhen weknowmore.
We are less likely to succeed when dealing with an unknown
situation than when dealing with a known one. The same
training is given to machines. In order for the machine to
make an accurate prediction, it needs to see an example.
Machines canpredictoutcomeswhengivensimilarexamples.
Like humans, the machine has difficulty predicting if it is fed
an unseen example [15].
The learningand inference processareatthecoreofmachine
learning. In order to learn, a machine must discover patterns
first. Data scientists must be careful when choosing which
data to feed to the machine. In solving a problem, a feature
vector consists of the attributes that go into theproblem.The
feature vector can be thought of as a subset of data that
addresses a particular issue. The machine simplifies the
reality and transforms it into a model using some fancy
algorithms. Consequently,amodelisdevelopedbydescribing
and summarizing the data.
Here are the points that summarize the lifecycle of Machine
Learning programs:
1. Question definition
2. Data collection
3. Data visualization
4. Algorithm training
5. Algorithm testing
6. Feedback collection
7. Algorithm refinement
8. Repeat steps 4-7 until satisfied with the results
9. Make a prediction based on the model
It applies new data sets to the algorithm once it becomes
adept at drawing the right conclusions.
5.4 Supervised Learning
To learn the relationship between inputs and outputs,
algorithms use training dataand humanfeedback.Ananalyst
can predict sales of cans by using marketing expenses and
weather forecasts, for example. When you know the output
data, you can use supervised learning. Predicting new data
will be the task of the algorithm.
Supervised learning can be divided into two categories:
1. Performing classifications
2. Performing regressions
5.5 Unsupervised Learning
It involves exploring input data without being given a clear
output variable (for example, to identify patterns from
customer demographic data). The algorithm will classify the
data for you if you do not know how to classify the data.
6. EXPERIMENTAL RESULTS
Fig-3: Home Screen

Fig-4: Upload File
Fig-5: Prediction Result
Fig-6: Performance Aanlysis
Fig-7: CropRecommendationfor Waterfall andTemperature
CONCLUSION
The latest technology can assist farmers in growing their
crops. Agriculturalists can be informed of accurate
predictionsofcropsinatimelymanner.Analysingagriculture
parameters has been done using a variety of Machine
Learning techniques. In a literature review, different
agricultural techniques are examined. Farmers can receive
personalized and relevant recommendations based on
parameters suchas production and season, resulting in good
crop yields.
REFERENCES
[1] Shreya S. Bhanose, Kalyani A. Bogawar (2021) “Crop
And Yield Prediction Model”, International Journal of
Advance Scientific Research and Engineering Trends,
Volume 1,Issue 1, April 2021
[2] Tripathy, A. K., et al.(2019) "Data mining and wireless
sensor network for agriculture pest/disease
predictions." Information and Communication
Technologies (WICT), 2019 World Congress on. IEEE.
[3] Ramesh Babu Palepu (2019) ” An Analysis of
Agricultural Soils by using Data Mining Techniques”,
International Journal of Engineering Science and
Computing, Volume 7 Issue No. 10 October.

[4] Rajeswari and K. Arunesh (2018) “Analysing Soil Data
using Data Mining Classification Techniques”, Indian
Journal of Science and Technology, Volume 9, May.
[5] A.Swarupa Rani (2020), “The ImpactofData Analyticsin
Crop Management based on Weather Conditions”,
International Journal ofEngineeringTechnology Science
and Research, Volume 4,Issue 5,May.
[6] Pritam Bose, Nikola K. Kasabov (2016), “Spiking Neural
Networks for Crop Yield Estimation Based on
Spatiotemporal Analysis of Image Time Series”, IEEE
Transactions On Geoscience And Remote Sensing.
[7] Priyanka P.Chandak (2017),” Smart Farming System
Using Data Mining”, International Journal of Applied
Engineering Research, Volume 12, Number 11.
[8] Vikas Kumar, Vishal Dave (2021), “KrishiMantra:
Agricultural Recommendation System”, Proceedings of
the 3rd ACM Symposium on Computing for
Development, January.
[9] SavaeLatu (2019), ”Sustainable Development : TheRole
Of Gis And Visualisation”, The Electronic Journal on
Information Systems in Developing Countries, EJISDC
38, 5, 1-17.
[10] Nasrin Fathima.G (2018), “Agriculture Crop Pattern
Using Data Mining Techniques”, International Journal of
Advanced Research in Computer Science and Software
Engineering, Volume 4, May.
[11] Ramesh A.Medar (2019), ”A Survey on Data Mining
Techniques for Crop Yield Prediction”, International
Journal of Advance Research in Computer Science and
Management Studies, Volume 2, Issue 9, September.
[12] L. Anand et al., “Development of machine learning and
medical enabled multimodal for segmentation and
classification of brain tumor using MRI images,”
Computational IntelligenceandNeuroscience,vol.2022,
pp. 1–8, Aug. 2022. doi:10.1155/2022/7797094
[13] Deshpande Radhika, Bhalekar Dipali, Mutkule Prasad,
Sanjay Pandhare, Nawale Akshay(2015) , “One Stop
Solution for Farmer Consumer”, Interaction, IJCA
Proceedings on National Conference on Advances in
Computing NCAC.
[14] Mutkule Prasad R., “Interactive Clothing based on IoT
using QR code and Mobile Application”, International
Journal of Scientific Research in Network Security and
Communication, vol. 6, issue-6, 2018.
[15] P. Mutkule and M. Ankoshe, “A survey on interactive
clothing based on IOT using QR code and Mobile
Application,” International Journal ofComputerSciences
and Engineering, vol. 6, no. 6, pp. 652–654, Jun. 2018.
doi:10.26438/ijcse/v6i6.652654

A Novel System for Recommending Agricultural Crops Using Machine Learning Approach

Recommended

More Related Content

Similar to A Novel System for Recommending Agricultural Crops Using Machine Learning Approach (20)

More from IRJET Journal (20)

Recently uploaded (20)

A Novel System for Recommending Agricultural Crops Using Machine Learning Approach