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

Volume: 04 Issue: 12 | Dec-2017 www.irjet.net p-ISSN: 2395-0072

Crop Recommendation System to Maximize Crop Yield using Machine

Learning Technique
Rohit Kumar Rajak1, Ankit Pawar2, Mitalee Pendke3 , Pooja Shinde4, Suresh Rathod5,
Avinash Devare6
123456Dept. of Computer Engineering, Sinhagad Academy of Engineering, Maharashtra, India

---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Agriculture in India plays a major role in 2. LITERATURE SURVEY
economy and employment. The common difficulty present
among the Indian farmers are they don’t opt for the proper The paper [1] states the requirements and planning
crop based on their soil necessities. Because of this needed for developing a software model for precision
productivity is affected. This problem of the farmers has farming is discussed. It deeply studies the basics of
been solved through precision agriculture. This method is precision farming. The author s start from the basics of
characterized by a soil database collected from the farm, precision farming and move towards developing a model
crop provided by agricultural experts, achievement of that would support it. This paper describes a model that
parameters such as soil through soil testing lab dataset. The applies Precision Agriculture (PA) principles to small, open
data from soil testing lab given to recommendation system farms at the individual farmer and crop level, to affect a
it will use the collect data and do ensemble model with degree of control over variability. The comprehensive
majority voting technique using support vector machine objective of the model is to deliver direct advisory services
(SVM) and ANN as learners to recommend a crop for site to even the smallest farmer at the level of his/her smallest
specific parameter with high accuracy and efficiency. plot of crop, using the most accessible technologies such as
SMS and email. This model has been designed for the
Key Words: Precision agriculture, recommendation scenario in Kerala State where the average holding size is
system, ensemble model, SVM, ANN, RANDOM TREE, NB- much lower than most of India. Hence this model can be
classifier. positioned elsewhere in India only with some
modifications.
1. INTRODUCTION The paper [2] makes a qualified cogitation of
assortment algorithms and their performance in yield
India is characterized by small farms. Over 75% of total prediction in precision husbandry. These algorithms are
land capitals within the country are less than 5 acres. Most implemented in a data set collected for several years in
crops are rain nourished, with just about 45% of the land yield prediction on soya bean crop. The algorithms used for
irrigated. As per some estimations, about 55% of total yield prediction in this paper are Support Vector Machine,
population of India depends on farming. In the US, because Random Forest, Neural Network, REPTree, Bagging, and
of heavy mechanization of agriculture, it is about 5%. Bayes. The conclusion drawn at the end is that bagging is
the best algorithm for yield prediction among the above
India is one of the biggest producers of agricultural stated algorithms since the error deviation in bagging is
products and still has very less farm productivity. minimum with a mean absolute error of 18985
Productivity needs to be increased so that farmers can get
more pay from the same piece of land with less labour. The paper [3] shows the importance of crop selection
Precision agriculture provides a way to do it. Precision and the factors deciding the crop selection like production
farming , as the name implies, refers to the applying of rate, market price and government policies are discussed.
precise and proper total of comment like pee , fertilizers, This paper proposes a Crop Selection Method (CSM) which
soil etc. at the proper time to the craw for increasing its solves the crop selection problem and improves net yield
productivity and increasing its yields. Not all precision rate of the crop. It suggests a series of crop to be selected
agriculture systems offer best results. But in agriculture it over a season considering factors like weather, soil type,
is important that the recommendations made are accurate water density, crop type. The predicted value of influential
and precise because in case of errors it may lead to heavy parameters determines the accuracy of CSM. Hence there is
material and capital loss. Many researches are being a need to include a prediction method with improved
carried out, in order to attain an accurate and efficient accuracy and performance.
model for crop prediction. Ensembling is one such
technique that is included in such research works. Among The paper [4] aims to solve the crucial problem of
these various machine learning techniques that are being selecting the classifiers for the ensemble learning. A
used in this field; this paper proposes a system that uses method to select a best classifier set from a pool of
the voting method to build an efficient and accurate model. classifiers has been proposed. The proposal aims to achieve
higher accuracy and performance. A method called SAD

© 2017, IRJET | Impact Factor value: 6.171 | ISO 9001:2008 Certified Journal | Page 950
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 12 | Dec-2017 www.irjet.net p-ISSN: 2395-0072

was proposed based on accuracy and classification soil. For crop growth to their possible, the soil must
performance. Using Q statistics, the dependency between provide acceptable environment for it. Soil is the anchor of
most relevant and accurate classifiers is identified. The the roots. The water holding capacity determines the
classifiers which were not chosen were combined to form crop's ability to absorb nutrients and other nutrients that
the ensemble. This measure is supposed to ensure higher are changed into ions, which is the form that the plant can
performance and diversity of the ensemble. Various use. Texture determines how porous the soil is and the
methods such as SA (Selection by Accuracy), SAD (Selection comfort of air and water movement which is essential to
by accuracy and Diversity) and NS (No selection) algorithm prevent the plants from becoming waterlogged. The level
were identified. Finally it is inferred that SAD works better of acidity or alkalinity (Ph) is a master variable which
than others. affects the availability of soil nutrients. The activity of
microorganisms present in the soil and also the level of
The paper [5] proposes various classification methods exchangeable aluminum can be affected by PH. The water
to classify the liver disease data set. The paper emphasizes holding and drainage determine the infiltration of roots.
the need for accuracy because it depends on the dataset Hence for the following reasons the above stated
and the learning algorithm. Classification algorithms such parameters are considered for choosing a crop.
as Naïve Bayes, ANN, ZeroR and VFI were used to classify
these diseases and compare the effectiveness, correction 3.2 Crop Prediction using Ensembling technique
rate among them. The performance of the models where
compared with accuracy and computational time. It was Ensemble is a data mining model also known as the Model
concluded that all the classifiers except naive bayes Combiners that combine the power of two or more models
showed improved predictive performance. Multilayer to attain better prediction, efficiency than any of its
perceptron show the highest accuracy among the proposed models could achieve alone. In our system, we use one of
algorithms. the most familiar Ensembling technique called Majority
Voting technique .In the voting technique any number of
The paper [6] tries to solve the problem of food base learners can be used. There should be at least two
insecurity in Egypt. It proposes a framework which would base learners. The learners are chosen in a way that they
predict the production, and import for that particular year. are capable to each other yet being complimentary also.
It uses Artificial Neural Networks along with Multi-layer Higher the competition higher is the chance of better
perceptron in WEKA to build the prediction. At the end of prediction. But it is necessary for the learners to be
the process we would be able to visualize the amount of complimentary because when one or few members make
production import, need and availability. Therefore it an error, the probability of the remaining members
would help to make decisions on whether food has to be correcting this error would be high. Each learner builds
further imported or not. itself into a model. The model gets trained using the
training data set provided. When a new data has to be
The soil datasets in paper [7] are analyzed and a
classified, each model predicts the class on its own. Finally,
category is predicted. From the predicted soil category the
the class which is predicted by majority of the learners is
crop yield is identified as a Classification rule. Naïve Bayes
voted to be the class label of the new sample.
and KNN algorithms are used for crop yield prediction. The
future work stated is to create efficient models using
various classification techniques such as support vector 3.3 Learners Used in the Model
machine, principal component analysis.
3.3.1 Support Vector machine
3. METHODOLOGY Support vector machines (SVM) is set of supervised
learning strategies used for classification, regression and
3.1 Dataset Collection outlier s discovery. it's a classification technique. Here, we
have a tendency to plot every information item as some
The dataset containing the soil specific attributes which
extent in n-dimensional house (where n is variety of
are collected from Polytest Laboratories soil testing lab,
options you have) with the worth of every feature being
Pune, Maharashtra, India. In addition, similar sources of
the worth of a selected coordinate. it's a classification
general crop data were also used from Marathwada
technique. during this algorithmic rule, we have a
University. The crops considered in our model include
tendency to plot every information item as some extent in
groundnut, pulses, cotton, vegetables, banana, paddy,
n-dimensional house (where n is variety of options you
sorghum, sugarcane, coriander. The number of examples
have) with the worth of every feature being the worth of a
of each crop available in the training dataset is shown. The
selected coordinate. A Support Vector Machine (SVM) is
attributes considered where Depth, Texture, Ph, Soil Color, discriminative classifier correctly bounded by a separating
Permeability, Drainage, Water holding and Erosion.
hyperplane. In alternative words, given labeled coaching
information (supervised learning), the algorithmic rule
The above stated parameters of soil play a major role in
outputs associate degree best hyperplane that categorizes
the crop's ability to remove water and nutrients from the
new examples. Support vector simple machine (SVM) may
© 2017, IRJET | Impact Factor value: 6.171 | ISO 9001:2008 Certified Journal | Page 951
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 12 | Dec-2017 www.irjet.net p-ISSN: 2395-0072

be a set of supervised learning strategies used for concerning cats, e.g., that they have fur, tails, whiskers and
classification, regression and outlier s uncovering . cat-like faces. Instead, they evolve their own set of
relevant characteristics from the learning material that
3.3.2 NAÏVE Bayes they process.

It is not single algorithm, but a clan of algorithmic rules . 3.3.4 Random Forest
All naive Bayes socio-economic classifiers adopts that the
value of a particular feature is independent of the value of Random forests square measure associate ensemble
any other feature, given the class variable. Naive Thomas learning methodology for classification, regression and
Bayes classifier could be a straightforward probabilistic different tasks, that operate by building a mess of call trees
classifier that works supported applying theorem (from at coaching time and outputting the category that's the
Bayesian statistics) with robust naive independence mode of the categories or mean prediction of the
assumptions. it's a classification technique supported individual trees. Random call forests correct for call tree
Bayes theorem with associate degree assumption of custom of over fitting to their coaching set. the primary
independence between predictors. In straightforward rule for random call forests was created by Tin Kam Ho
terms, a Naive Thomas Bayes categoryifier assumes that victimization the random mathematical space
the presence of a specific feature in an exceedingly class is methodology, which, in Ho's formulation, could be a
dissimilar to the presence of the other feature. as an thanks to implement the "stochastic discrimination"
example, a fruit could also be thought of to be associate approach to classification projected by Eugene Kleinberg.
degree apple if it's red, round, and concerning a pair of associate addition of the rule was developed by Leo
inches in diameter. Even if these features depend on each Breiman and Adele bargainer, and "Random Forests" is
other or upon the existence of the other features, a naive their trademark. The addition combines Breiman's
Bayes classifier would consider all of these properties to "bagging" plan and random choice of options, introduced
independently contribute to the probability that this fruit 1st by Ho and later severally by Amit and Geman so as to
is an apple. These Learners predict the class label for each construct a group of call trees with controlled variance.
of the training data set. The class label that is predicted by though random forests are naturally designed to figure
the majority of the models is voted through the majority solely with third-dimensional information, it's been shown
voting technique and the class label of the training data set that one may use them for random objects with the
is decided. From the ensembled models the rules are employment of solely pairwise similarities between
generated. objects.

3.3.3 Multi-layer Perceptron (Artificial Neural Random Forest could be a trademark term for associate
Network) ensemble of call trees. Random Forest is assortment of call
trees (so called Forest). To classify a replacement object
Multi Layer perceptron (MLP) could be a feed forward supported attributes, every tree offers a categoryification
neural network with several layers between input and end and that we say the tree votes for that class.
product layer. Feed forward implies that knowledge
period in one direction from input to output layer. MLPs 3.4 Rules induced from the Model
square measure typically used for pattern classification,
recognition, anticipation and approximation. Associate The rule below demonstrates an example of the proposed
ANN relies on a set of connected units or nodes referred to recommendation system.
as artificial neurons (analogous to biological neurons in IF ph is mild alkaline
associate animal brain). every association (synapse) AND depth is above 90
between neurons will transmit an indication from one to a AND water holding capacity is LOW
different. The receiving (postsynaptic) somatic cell will AND drainage is moderate
method the signal(s) then signal neurons connected to AND erosion is LOW
that. Artificial neural networks (ANNs) or connectionist THEN PADDY
systems square measure computing systems galvanized by The IF part of the rule states the soil specifications needed
the biological neural networks that represent animal for the cultivation of the recommended crop which is
brains specified in the THEN PART of the rule.

In Neural Networks some nodes use a nonlinear activation

4. CONCLUSION
operate that was developed to model the frequency of
action potentials, or firing, of biological neurons. for
instance, in image recognition, they could learn to spot Our work would help farmers to increase productivity in
pictures that contain cats by analyzing example pictures agriculture, prevent soil degradation in cultivated land,
that are manually labelled as "cat" or "no cat" and and reduce chemical use in crop production and efficient
exploitation the results to spot cats in different pictures. use of water resources. Our future work is aimed at an
they are doing this with none a priori information

© 2017, IRJET | Impact Factor value: 6.171 | ISO 9001:2008 Certified Journal | Page 952
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 04 Issue: 12 | Dec-2017 www.irjet.net p-ISSN: 2395-0072

improved data set with large number of attributes and

also implements yield prediction.

REFERENCES

[1] Satish Babu (2013), A Software Model for Precision

Agriculture for Small and Marginal Farmers , at the
International Centre forFree and Open Source
Software (ICFOSS) Trivandrum, India.

[2] Anshal Savla, Parul Dhawan, Himtanaya Bhadada,

Nivedita Israni, Alisha Mandholia , Sanya Bhardwaj
5 , Survey of classification algorithms for
formulating yield prediction accuracy in precision
agriculture', Innovations in Information,Embedded
and Communication systems (ICIIECS).

[3] Rakesh Kumar, M.P. Singh, Prabhat Kumar and J.P.

Singh 5 , Crop Selection Method to Maximize
Crop Yield Rate using Machine Learning Technique ,
International Conference on Smart Technologies and
Management for Computing, Communication,
Controls, Energy and Materials (ICSTM).

[4] Liying Yang , Classifiers selection for ensemble

learning based on accuracy and diversity Published
by Elsevier Ltd. Selection and/or peer-review under
responsibility of [CEIS].

[5] A.T.M Shakil Ahamed, Navid Tanzeem Mahmood,

Nazmul Hossain, Mohammad Tanzir Kabir, Kallal Das,
Faridur Rahman, Rashedur M Rahman (2015) ,
Applying Data Mining Techniques to Predict Annual
Yield of Major Crops and Recommend Planting
Different Crops in Different Districts in Bangladesh ,
(SNPD) IEEE/ACIS International Conference.

[6] Aymen E Khedr, Mona Kadry, Ghada Walid (2015),

Proposed Framework for )mplementing Data Mining
Techniques to Enhance Decisions in Agriculture Sector
Applied Case on Food Security Information Center
Ministry of Agriculture, Egypt , )nternational

[7] Monali Paul, Santosh K. Vishwakarma, Ashok Verma

(2015), Analysis of Soil Behaviour and Prediction of
Crop Yield using Data Mining Approach , )nternational
Conference on Computational Intelligence and
Communication Networks

© 2017, IRJET | Impact Factor value: 6.171 | ISO 9001:2008 Certified Journal | Page 953