International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 08 Issue: 04 | Apr 2021 www.irjet.net p-ISSN: 2395-0072

Crop and Fertilizer Recommendation System using

Machine Learning
Palaniraj A1, Balamurugan A S2, Durga Prasad
R3,Pradeep P4

1Assistant Professor, Department of Information Technology, Panimalar Institute of Technology,

Chennai, India
2,3,4Student, Department of Information Technology, Panimalar Institute of Technology, Chennai,
India.
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract—India being an agriculture country, its economy wetland it is better to go for wheat and sugarcane. There are
predominantly depends on agriculture yield growth and 15 agro-climatic regions in India these regions are divided on
the bases of a type of the land. Each agro-climatic regions can
agroindustry products. Data Mining is an emerging research
grow some specific crops. Based on that we need to suggest
field in crop yield analysis. Yield prediction is a very
the farmer that which crop is best among those crops which
important issue in agricultural. Any farmer is interested in belong to those climatic regions. Achieving the maximum
knowing how much yield he is about to expect. Analyze the crop at minimum yield is the ultimate Aim of the project.
various related attributes like location, pH value from which Early detection of problems and management of those
alkalinity of the soil is determined. Along with it, percentage problems can help the farmers for better crop yield. Crop
of nutrients like Nitrogen (N), Phosphorous (P), and yield prediction is the important research which helps to
Potassium (K) Location is used along with the use of third- secure food. For the better understanding of the crop yield,
party applications like APIs for weather and temperature, we need to study of the huge data with the help of machine
type of soil, nutrient value of the soil in that region, amount learning algorithm so it will give the accurate yield for that
of rainfall in the region, soil composition can be determined. crop and suggest the farmer for a better crop. Improving the
All these attributes of data will be analyzed, train the data quantity of the crop is the key goal of precision agriculture
with various suitable machine learning algorithms for means obtaining a better understanding of the crop using the
creating a model. The system comes with a model to be information technology methods. The main goal of precision
precise and accurate in predicting crop yield and deliver the agriculture is profitability and sustainability. From ancient
end user with proper recommendations about required times agriculture has become the backbone of our country.
fertilizer ratio based on atmospheric and soil parameters of Nowadays climatic conditions vary very often. So, it is hard
the land which enhance to increase the crop yield and to grow crops by understanding weather conditions. We
need to use some technology to find or understand the crop
increase farmer revenue.
details and guide the farmers to grow crops accordingly and
moreover fertilizer also one of the major factors to grow
Key Words: Crop yield prediction, Support Vector
crops accordingly. If fertilizer is used more or less in the field
Machine. the soil may lose it fertility and crop may not give the
expected yield. So, fertilizer also becomes the major factor in
it. mostly understanding the temperature conditions is much
1. INTRODUCTION necessary for India because we can improve the Indian
economy with the help of the crop prediction because it
INDIA is a highly populated country and randomly change in plays a major role in the Indian economy. Generally, machine
the climatic conditions need to secure the world food learning algorithms will predict the most efficient output of
resources. Framers face serious problems in drought the yield. Previously yield is predicted on the bases of the
conditions. Type of soil plays a major role in the crop yield. farmers prior experience but now weather conditions may
Suggesting the use of fertilizers may help the farmers to change drastically so they cannot guess the yield. so,
make the best decision for their cropping situation. The technology can help them to predict the yield of the crop
number of studies Information and Communication weather to go for that crop or no. machine learning model
Technology (ICT) can be applied for prediction of crop yield. will understand the pattern of the crop and yield based on
By the use of Data Mining, we can also predict the crop yield. the several conditions and predicts the yield of the area in
By fully analyze the previous data we can suggest the farmer which he is going to crop.
for a better crop for the better yield. For the better yield we
need to consider soil type and soil fertility things and also
one of the major factors rainfall and groundwater availability
if it is dry land it is better to go for cash crops and if is

© 2021, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 319
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 08 Issue: 04 | Apr 2021 www.irjet.net p-ISSN: 2395-0072

LITERATURE SURVEY 10].Jintrawet, A. et al. (2008, May) used SVR model for
crops like rice to predict the yield where the model was
Ananthara,M.G.etal.(2013,February)proposed divided into three steps- predicting the soil nitrogen weight
prediction model for datasets pertaining to agriculture which followed by prediction of rice stem weight and rice grain
iscalledasCRYalgorithmforcropyieldusingbeehiveclusteringte weight respectively. Their factors were solar radiation,
chniques.Theyconsideredparametersnamelycroptype, soil temperature and precipitation along with those three steps
type, soil pH value, humidity and crop sensitivity. Their
analysis was mainly in paddy, rice and sugarcane yields in [11].Miniappan,N.etal.(2014,August)usedartificialneuraln
India.Their proposed algorithm was then compared with. etworkinmodelling multi-layer perceptron model with 20
C&R tree algorithm and it outperformed well with an hidden layers for prediction wheat yield which considered
accuracy of 90 percent factors likes unlight, rain, frost and temperature
[2]. Awan, A. M. et al. (2006, April) built a new, smart [12]. Manjula, A et al.built crop selection and to predict
framework focused on farm yield prediction the yield which considered various indexes like vegetation,
clusteringkernelmethodologyandtheyconsideredparametersli temperature and normalized difference vegetation as factors.
keplantation, latitude, temperature and precipitation of They distinguished between climate factors and agronomic
rainfall in that latitude. They had experimented weighted k- factors and other disturbances caused in the prediction for
means kernel method with spatial constraints for the analysis better understanding
of oil palm fields
[13]TamilNadu, India.They have considered actors like
[3]. Chawla, I. et al. (2019, August) used fuzzy logic for soil, temperature, sunshine, rainfall, fertilizer, paddy, and
crop yield prediction through statistical time series models. type of pest used and other factors like pollution and season
They considered parameters like rainfall and temperature for
[14]. Verma, A. et al. (2015, December) used
prediction.Their prediction was classification with levels
classification techniques like NaïveBayes,K N
‘good yield’ , ‘very good yield’
algorithm for crop prediction on soil datasets which
[4]. Chaudhari, A. N. et al.(2018,August)used three constituted nutrients of soil like zinc, copper,
algorithms namely clusterin gk-means, A priori and Bayes manganese, pH, iron, Sulphur, Phosphorous,
algorithm, then they hybridized the algorithm for better Potassium, nitrogen, and Organic Carbon
efficiency of yield prediction and they considered parameters
[15].Kalbande,D.R.et al. (2018) used support
like Area, Rainfall, Soil type and also their system was able to
vector regression, multi polynomial regression and
tell which crop is suitable for cultivation based on the
random forest regression for prediction of cornyield
mentioned features
and evaluated the models using metrics like errors
[5]. Gandge, Y.(2017, December) used many machine namely MAE,RMSE and R-square values
learning algorithms for different crops. They studied and
[16].Rahman,R.M.etal. (2015, June) used mainly
analyzed which algorithm would be suitable for which crop.
clustering techniques for cropyield prediction. The
They have used K-means, Support vector Regression, Neural
paper explained the analysis of majorcrops in
Networks, C4.5 Decision tree, Bee-Hive Clustering, etc. The
Bangladesh and divided the variables into
factors implying were soil nutrients like N, K, P and soil ph.
environmental and biotic variables.The algorithms
[6]. Armstrong, L. J. et al.(2016, July) used ANNs for the applied were line arregression, ANN,and KNN
prediction of rice yield in the districts of Maharashtra, India. approach for classification
They considered climatic factors namely (considering range)
[17]. Hegde, M. et al. (2015, June) used multiple
temperature, precipitation and reference cropevapo
linear regression and neuro fuzzy systems for
transpiration. The records were collected from Indian
predicting cropyield by taking biomass, soil water,
Government repository from 1998 to2002 [7]. Tripathy, A. K.
radiation and rainfall as input parameters for the
et al. (2016, July) were same authors who used support
research and their majorly concentrated crop was
vector machines to predict the rice crop yield with same
wheat
features as the previous paper mentioned
[18]. Sujatha, R., & Isakki, P.(2016, January) used
[8].Petkar,O. (2016, July) were also the same authors who
classification techniques like ANN, j48,Naïve Bayes,
applied for SV Mandneural networks for rice cropyield
Random Forest and Support vector Machines. They
prediction proposed a new decision system which is an
have also included both climatic parameters and soil
interface to give the input and get the output
parameters as features in their modelling
[9].Chakrabarty,A.etal.(2018,December)nalyzed crop
[19]. Ramalatha, M.et al. (2018, October) used a
prediction in the country of Bangladesh where they majorly
hybrid approach of combining K-means clustering
cultivate three kinds of rice, Jute, Wheat, and Potato. Their
and classification based on modified KNN
research used a deep neural network where the data had
approach.The data was collected from TamilNadu,
around 46 parameters into their consideration. Few of them
India where the majorly concentrated crops were
were soil composition, type of fertilizer ,type of soil and its
rice, maize, Ragi, Sugarcane, and Tapioca
structure ,soil consistency, reaction and texture

© 2021, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 320
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 08 Issue: 04 | Apr 2021 www.irjet.net p-ISSN: 2395-0072

[20].Singh,C.D.etal.(2014,January)developedanap Step-4: Now,the prediction test results are used to

plicationtoadvisecropswhichworkson selected
districts of Madhya Pradesh, India. The user would measure the accuracy.
give input cloud cover, rainfall, temperature,
observed yield in the past and the system would To separate the two classes of data points, there are
predict the yield and Depending on the trigger values many possible hyperplanes that could be chosen. Our
set, the crop will be labeled and obtain the results. objective is to find a plane that has the maximum margin,
that is the maximum distance between data points of
both classes. Maximizing the margin distance provides
some reinforcement so that future data points can be
METHODOLOGY classified with more confidence.
Hyperplanes are decision boundaries that help
classify the data points. Data points falling on either side
of the hyperplane can be attributed to different classes.
Also, the dimension of the hyperplane depends upon the
number of features. If the number of input features is 2,
then the hyperplane is just a line. If the number of input
features is 3, then the hyperplane becomes a two-
dimensional plane. It becomes difficult to imagine when
the number of features exceeds 3.
Support vectors are data points that are closer to the
hyperplane and influence the position and orientation of
the hyperplane. Using these support vectors, we
maximize the margin of the classifier. Deleting the
support vectors will change the position of the
hyperplane. These are the points that help us build our
SVM.fig.3.

Fig.2.Processchartoftheresearchproject

A. Pre-processing
For the given data set, there are few values missing in
the data set, the missing values could lead to false
prediction results. The missing values are removed
using the python. It is also important to remove the
replicated data from the data set to fetch the accurate
results. Normalization is used to scale the data to a
specific range.

B. Support Vector Machine: Fig.3.Support Vector Machine

The Support Vector Machine is one of the supervised

machine learning algorithm mainly used in classification C. Output:
and regression problems. We plot each data item as a The Support Vector Machine Used in this project gives
point in n-dimensional space with the value of each the accuracy over 90.01%, Besides the accuracy of the
feature being the value of a particular coordinate. SVM is (mean accuracy=0.950; mean AUC=0.934; mean
F1 score=0.903)The user or the farmer can enter the
Step1:the total training set is again divided into two following details over the web application to get the
different sets.(train and holdout) prediction as depicted below in the fig 4
Step2:train thes elected base models with first
part (train).

Step-3:Test them with these cond part.(holdout)

© 2021, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 321
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 08 Issue: 04 | Apr 2021 www.irjet.net p-ISSN: 2395-0072

2019TwelfthInternationalConferenceonContemp
oraryComputing (IC3)(pp. 1-6).IEEE.
[5] Bhosale, S. V., Thombare, R. A., Dhemey, P. G., &
Chaudhari, A. N.(2018, August). Crop Yield
Prediction Using Data Analytics and Hybrid
Approach. In 2018Fourth International
Conferenceon Computing Communication Control
and Automation(ICCUBEA) (pp.1-5).IEEE.
[6] Gandge,Y.(2017,December).A study on various data
mining techniques for crop yield prediction. In 2017
Fig.4.InterfaceofWebApp International Conference on Electrical, Electronics,
Communication, Computer, and Optimization
Techniques(ICEECCOT)(pp.420-423).IEEE.
CONCLUSION AND FUTUREWORK [7] Gandhi, N., Petkar, O., & Armstrong, L. J. (2016, July).
The core strategy of this project is to predict the crop Rice crop yield prediction using artificial neural
based on the soil nutrient content and the location networks. In 2016 IEEE Technological Innovations in
ICT for Agriculture and Rural Development (TIAR)
where the crop is growing. This system will help he
(pp.105-110).IEEE.
farmers to choose the right crop for their land and to
give the suitable amount of fertilizer to produce the [8] Gandhi, N., Armstrong, L. J., Petkar, O., & Tripathy, A.
maximum yield. The Support Vector Machine algorithm K. (2016, July).Rice crop yield prediction in India
helps to predict the crop the precisely based on the using support vector machines.
pre-processed crop data. This system will also help the In201613thInternationalJointConferenceonComput
new comers to choose the crop which will grow in erScienceandSoftwareEngineering (JCSSE) (pp.1-5).
their area and produce them a good profit. A decent IEEE.
amount of profit will attract more people towards the [9] Gandhi,N.,Armstrong,L.J.,&Petkar,O.(2016,July).Prop
agriculture. Also, the crop growth is based on the osed decision support system (DSS) for Indian rice
climate conditions in the particular area and the crop yield prediction. In2016 IEEE Technological
seasonal monsoons happens now are unpredictable, Innovations in ICT for Agriculture and Rural
hence it is easy for the farmers when the prediction Development (TIAR)(pp.13-18). IEEE.
result is also based on the climatic conditions. Live [10] Islam, T., Chisty, T. A., & Chakrabarty, A. (2018,
weather prediction will also help the users to predict December). A Deep Neural Network Approach for
the crop water needs and also it will help the farmers Crop Selection and YieldPrediction[1] in Bangladesh,
to decrease the crop damage due to the rain or In 2018 IEEE Regio 10 Humanitarian Technology
drought. Conference(R10-HTC) (pp. 1-6). IEEE.
[11] Jaikla, R., Auephanwiriyakul, S., & Jintrawet, A. (2008,
REFERENCE May). Riceyield prediction using a support vector
regression method. In 2008
5thInternationalConferenceonElectricalEngineering/
[1] “Kaglee.com.”[Online].Available:https://Kaglee.com/ Electronics,Computer,Telecommunications and
[2] Ananthara, M. G., Arunkumar, T., & Hemavathy, Information Technology(Vol.1,pp.29-32).IEEE.
R. (2013, February).CRY—an improved [12] Kadir, M. K. A., Ayob, M. Z., & Miniappan, N. (2014,
cropyield prediction model using beehive August). Wheatyield prediction: Artificial neural
clustering approach for agricultural datasets .In network based approach. In 2014 4thInternational
2013InternationalConferenceonPatternRecogniti Conference on Engineering Technology and
on,Informatics and Mobile Engineering(pp.473- Technopreneuship(ICE2T)(pp.161-165).IEEE.
478).IEEE. [13] Manjula, A., & Narsimha, G. (2015, January). XCYPF:
[3] Awan, A. M., & Sap, M. N. M. (2006, April). An A flexible and extensible framework for agricultural
intelligent system based on kernel methods for Crop Yield Prediction. In 2015IEEE 9th International
cropyield prediction. In Pacific-AsiaConference Conference on Intelligent Systems and
on Knowledge Discovery and Data Mining (pp. Control(ISCO)(pp.1-5).IEEE.
841-846).Springer,Berlin,Heidelberg [14] Mariappan, A. K., & Das, J. A. B. (2017, April). A
[4] Bang, S., Bishnoi, R., Chauhan, A. S., Dixit, A. K., & paradigm for riceyield prediction in Tamilnadu. In
Chawla, I. (2019,August). Fuzzy Logic based 2017 IEEE Technological Innovations in ICT for
Crop Yield Prediction using Temperature and Agriculture and Rural Development(TIAR)(pp.18-
Rainfall parameters predicted through 21).IEEE.
ARMA,SARIMA,andARMAXmodels. In [15] Paul,M.,Vishwakarma,S.K.,&Verma,A.(2015,Decembe

© 2021, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 322
International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 08 Issue: 04 | Apr 2021 www.irjet.net p-ISSN: 2395-0072

r).Analysisofsoilbehaviourandpredictionofcropyield
usingdatamining approach. In 2015 International
Conference on Computational Intelligence and
CommunicationNetworks(CICN)(pp.766-771).IEEE.
[16] Shah, A., Dubey, A., Hemnani, V., Gala, D., & Kalbande,
D. R. (2018).Smart Farming System:CropYield
Prediction Using Regression

© 2021, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 323