IRJET- Analysis of Crop Yield Prediction using Data Mining Technique to Predict Annual Yield of Major Crops

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 12 | Dec 2018 www.irjet.net p-ISSN: 2395-0072
© 2018, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 1460
ANALYSIS OF CROP YIELD PREDICTION USING DATA MINING
TECHNIQUE TO PREDICT ANNUAL YIELD OF MAJOR CROPS
B. Devika¹, B. Ananthi²
¹Research scholar, M.Phil. Computer Science, Vellalar College for Women, Erode12.
²Associate Professor and Head, Department of Computer Science, Vellalar College for Women, Tamilnadu, India.
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - India is generally an agricultural country.
Agriculture is the single most importantprovidertotheIndian
economy. Agriculture crop production depends on the season,
organic, and monetary cause. The prognostication of
agricultural yield is challenging and pleasing task for every
nation. Nowadays, Farmers are hostile to produce the yield
because of erratic climatic changes and scarcity of water
resource. The main objective is collecting agricultural data
which can be stored and analyzed for useful crop yield
forecasting. To predict the crop yield with the help of data
mining technique, advanced methods can be introduced to
predict crop yield and it also helps the farmer to choose the
most suitable crop, thereby improving the value and gain of
the farming area.
Key Words: Data Mining, Classification, Crop Yield,
Accuracy, K-Nearest Neighbor (KNN), LinearRegression
1. INTRODUCTION
Data Mining is the process of extract helpful and
significant information from huge sets of data. Data Mining
in agriculture field is a comparatively novel research field.
Yield prediction is a very important agricultural problem.
Any farmer is interested in knowing how much yield he is
concerning to be expecting. In the earlier period, yield
prediction was performing by considering farmer's
experience on particular field and crop. In any of Data
Mining actions the training data is to be collected from past
data and the gathered data is used in terms oftraining which
has to be exploited to study how to categorize future yield
predictions. Crop models and decision tools are more and
more used in agricultural field to improve production
efficiency. The combination of higher technology and
agriculture to improve the production of crop yield is
becoming more interesting newly. Due to the rapid
development of new high technology, crop models and
predictive tools might be predictable to become a crucial
element of agriculture.
Crop yield is a combined bio-socio-system
comprised of complex interaction among the soil,theair,the
water, and the crops grown in it, where a comprehensive
model is necessary which are possible onlythroughclassical
engineering expertise. As define by theFoodandAgriculture
of the United Nations, crop forecasting is the art of predict
crop yields and production before the harvest in fact takes
place, typically a couple of months in advance. Crop
forecasting philosophy is based on various kinds of data
collected from different sources: meteorological data, agro-
meteorological, soil, remotely sensed, agricultural statistics.
Based on meteorological andagronomicdata,several indices
are derived which are deemed to be relevant variables in
determining crop yield, for instance crop water satisfaction,
surplus and excess moisture, average soil moisture, etc.
Linear Regression model characterizes the mathematical
relationships intrinsic to the data set from previous
experiments. This methodcanproduceresultsundervarious
situations assuming extensive information used to expand
and test the model. Though, in agricultural data,information
is rather sparse and incomplete. Because of this limitation,
the linear regression approach is the common approach for
predicting yield across large area.
The most investigate statistical crop-yield-weather
model are multivariate regression models. Data mining
technique aim at finding those patternsorinformationinthe
data that are together valuable and interestingtothefarmer.
A common specific problem to occur is yield prediction.
2. LITERATURE REVIEW
Rajshekhar Borate etc.al [2016] describes and
gave the details us for list of used methods, In India there are
dissimilar agriculture crops production and those crops
depends on the several kindoffactorssuchasenvironmental
science, economy and also the geographical factorscovering
such methodologies and methods on historic yield of
dissimilar crops, it is possible to get info or data which can
be supportive to farmers and government organizations for
creation well decisions and for make better ruleswhichhelp
to increased production. In this article, our effort is on
application of data mining techniques which isusetoextract
information from the agricultural records to estimate better
crop yield for main crops in main districts of India. In our
project we found that the precise prediction of dissimilar
specified crop yields across different districts will help to
farmers of India. From this Indian farmers will plant
different crops in different distr [8].
Ramesh, D., and VishnuVardhan, B.,Agrarian
et.al [2015] discussed a several subdivisioninIndia isfacing
rigorous problem to make the most of the crop productivity.
More than 60 out of a hundred the crop still depends on
monsoon rainfall. Current growths in Information
Technology for agriculture field have developed an
interesting research area to forecast the crop yield. The
problematic of yield prediction is a major problem that
remains to be solved based on accessible data. Data mining
methods are the better selections for this purpose. Different
Data Mining methods are used and evaluated in agriculture

for approximatingtheupcomingyear'scropproduction. This
paper presents a brief analysis of crop yield predictionusing
Multiple LinearRegression(MLR)methodandDensitybased
clustering technique for the particular region i.e. East
Godavari district of Andhra Pradesh inIndia.Inthispaper an
effort is made in command to know the region precise crop
yield analysis and it is processed by applying both Multiple
Linear Regression method and Density-based clustering
method. These models were experimented in respect of all
the districts of Andhra Pradesh, then the procedure of
evaluation is passed out with only East Godavari district of
Andhra Pradesh in India [1].
Veenadhari,S.,Bharat Misra,DSingh et.al[2011]
discussed that the data mining extraction of unseen
predictive information from huge records,isa powerful new
technology with great potential to help companies focus on
the most significant data in their Data warehouses. Data
mining tools predict upcoming trends and performance and
growth, allowing businesses to make proactive, knowledge-
driven decisions. Though these methods are plausible,
theoretically well created, and perform well on extra or less
artificial test data sets, they depend on their skill to make
sense of real-world data. This article gave us a detail project
that is smearing a range of machine learning plans to
problems in agriculture and horticulture. They briefly
surveyed some of the techniques emerging from machine
learning study, define a software workbenchfortesting with
a variability of methods on real-world data sets, and a
learning of dairy herd managementinthatcullingruleswere
inferred from a medium-sized record of herd information.
They also defined a range of machine learning plans to
problems in agriculture and horticulture. There is a rising
number of applications of data mining methods in
agriculture and a rising amount of data that are presently
available from several resources. This is relatively a novel
research field and it is expected to grow in the upcoming.
There is a lot of effort to be done on this emerging and
interesting study field. The multidisciplinary method of
integrating computer science with agriculture will help in
predicting managing agricultural crops effectively [10].
Dakshayini Patil etc.al [2017] describes and
discover the list of methods and techniques which are used
Rice crop creation assumes an imperativepartinsustenance
safety of India, contributing over 40% to general yield
generation. High harvestgenerationisreliantonappropriate
climatic situations. Inconvenient regular atmosphere
conditions, for example, low precipitation or temperature
extremes can drastically diminish edit yield. Rising well
plans to foresee edit efficiency in several climatic conditions
can help rancher and different partners in vital basic
leadership as far as agronomy and yield result. This article
reports utilization of many information mining approaches
will anticipate rice trim yield forMaharashtra state,India. To
this review, 27 regions of Maharashtra were picked on the
establishment of accessible information from openly
available Indian Administration records with different
atmosphere and yield limitations. This surveys thetechnical
achievements in the field of Rice crop yield prediction.
Discuses methodology, comprehensive survey of many
proposed approaches to predict rice crop yield and
applications. It also discusses various data mining methods
used for prediction of crop yield for rice. Rising better plans
to foresee crop productivity in various climatic conditions
can help farmer and different partners in essential basic
leadership as far as agronomy and product decision [2].
Ramesh A. Medar and Vijay. S. Rajpurohit et.al
[2014] presented a Precision agriculture (PA) and
information technology (IT) are closely interwoven. The
former frequently refers to the application of nowadays’
technology to agriculture. Due to the use of sensors and GPS
technology, in today’s agriculture several data are collected.
Creation use of those data via IT often leads to dramatic
improvements in efficiency. For this purpose, the challenge
is to change these raw data into useful data. This paperdeals
with suitable modeling methods for those agricultural data
where the objective is to uncover the surviving patterns. In
specific, the use of feed-forward back propagation neural
networks will be evaluated and suitable parameters will be
projected. In consequence, yield prediction is allowedbased
on cheaply obtainable site data. In this prediction, economic
or environmental optimization of, e.g., fertilization can be
passed out. Due to the rapidly advancing technology in the
last few decades, ever more of our everyday life has been
changed by information technology. Data access, once
cumbersome and slow, has been turned into “data at your
fingertips” at high speed. Technological breakthroughshave
been made in industry and services as well as in agriculture
[9].
V. Leemans, M.-F. Destain et.al [2004] describes
that the suggested Fresh market fruitslikeapplesaregraded
into quality groups according to their size, color and shape
and to the attendance of defects.Thetwofirstqualitycriteria
are actually automatic on industrial graders, then fruits
classifying according to the presence of faults is not yet
efficient and so remains a manual operation, repetitive,
luxurious and not reliable. The classifying of apples using
machine vision can be arbitrarily separated into four steps:
the images acquisition, their segmentation, their
interpretation and in conclusion the fruit classification. This
paper presents the three former points on the basis of a
literature review, the research outcomes being absorbed on
the last point: having extracted data from images acquired
on fruits, the paper definesa classifyingtechniquewhichwas
implemented on an existing machine and tested on Jon gold
apples (bi-color fruits).The first step consists of acquiring
images of the surface of the fruit, though it goes through the
classifying machine. In order togradeapples,two necessities
have to be met: the images should cover the entire surface of
the fruit; a high contrast has to be created among the defects
and the healthy tissue, while maintaining a low variability
for the healthy tissue [4].
Roger J. Brooks, Mikhail. Semenov, Peter D.
Jamieson et.al [2001] presented simpler meta-model,
which produced very similar yield predictions to Sirius of
potential and water-limited yields at twolocationsinthe UK,

Roth Amsted and Edinburgh. This greatly increases the
understanding of the nature and consequences of the
relationships implicit within Sirius. The study showed that
the reply of wheat crops to climate could be explain using a
few simple relationships. The meta-model aggregate the
three main Sirius components, the computation of leaf area
index, the soil water balance model and the
evapotranspiration calculation, into simpler equations.
These results in a obligation for calibration of fewer model
parameters and means that weather variables can be
provide on a monthly rather than a daily time-step, because
the meta-model can use cumulative values of weather
variables. As a result the meta-model is a valuable tool for
regional impact assessments when detailed input data are
usually not available. As the meta-model was developed
from the analysis of Sirius, rather than from statistical fitting
of yield to weather data, it must do well forotherlocationsin
Great Britain and with different management scenarios [7].
Prof. M.S .Prasad Babu, N.V.Ramana Murty,
S.V.N.L.Narayana et.al [2010] describes that the tomato is
now the most widely grown vegetable crop in World. It is
grownup through the world in farm gardens, small home-
gardens, and by market gardeners for fresh consumption as
well as for processing purposes. This Tomato crop skilful
advisory system is intended at a collaborative venture with
eminent Agriculture Scientist and Experts in the area of
Tomato Plantation with an excellent team of computer
Engineers, computer programmer and creators. This Expert
System contains two main parts one is Tomato Info System
and the other is Tomato Crop Expert System where in Data
system, the user can get all the static information about
different species, Illnesses, Symptoms, chemical controls,
Preventions, Pests, Virus of Tomato fruits and plants. In
Advisory System, the user is having a communication with
the expert system online; the user has to answer the
questions asked by the Expert System. Depends on the reply
by the operator the expert system decides the disease and
shows its control measure of disease.ThisTomatoCropData
Expert System deals with different varieties ofTomatoCrop,
Identification of various diseases usually chances to tomato
crop based on the symptoms [6].
3. SYSTEM METHODOLOGY
3.1 ARCHITECTURE OF CROP YIELD PREDICTION
The crop yield prediction includes repeatedly all
essential parameters that are needed for the well yield of
crop. This improves the classification outcomes of the crop
yield. All the essential parameters are thought-about as
inputs. In common, one in all the issues faced with in the
prediction method is that almost all of the required
parameters that are essential to consider for the exact
prediction are not consider. It decreases the efficiencyofthe
anticipated outcomes which in turn leads to lack of proper
forecasting of the crop harvest its additionally tougher to
predict the improved predict the improved range of input
parameters that are to be considered in the prediction
procedure.
Crop prediction is that the art of predicting crop
yields and manufacture before the yield really takes place.
Before harvest prediction was done by considering the
farmer’s knowledge on a selected field and crop. This work
presents a system that uses data processing strategies so as
to predict the analyzed datasets. The anticipated sort can
specify the yielding of crops.
Architecture is a system that unites its parts or
components into a coherent and purposeful complete. The
crop information base consists of farm data like crop
varieties, crop year, area and seasonal parameter like Khrif,
rabbi and summer crops. The knowledge-basedadditionally
contains of zones furthermore district information,
ecological parameter like extreme and lowest temperature
value and average precipitation.
The crop yield prediction model that includes
associate input module that is in charge of taking input
from the farmer. The input module includes crop name,
land area, crop year and prediction tons The feature
selection model is in charge offset. Selection of associate
attribute from crop particulars. The crop yield prediction
model used to predict the yield. Once feature selection, the
data go to classification rule for grouping similar contents
Climate Crop Area
data data Parameters
Fig 3.1 Architecture of crop yield Prediction
Data Mining
Data Collections
Yields
Production (in
tons)
Classification Algorithm
(Linear Regression)
Weather
Statement
Cultivated AreaCrop

Climate data and crop parameters used to predict
crop growth can be predicted. Then prediction rules are
going to be applied to the output of classifying crop
particulars in terms of crop name, season and total yield
details.
3.2 IMPLEMENTATION
Regression analysis is used to analyze and
determine the affiliation between response variable and
explanatory variable. The variables considered for analysis
during this analysis work are annual prediction, area under
cultivation. Crop yield may be dependent on variable which
depends on all these ecological factors.
Linear Regression
A Linear regression methodology that’s used to
analyze a response variable that alterationswiththevalueof
the interference variable. A way of predicting the value of a
response variable from a given value of the explanatory
variable is also referred to as prediction.Theleast-square fit,
that is capable of fitting each linear additionally as
polynomial relations, is that the most typically used linear
regression. The method of applyingmodel estimatetovalues
outside the domain of the first knowledge is thought as
extrapolation. A linear regression model is computed to
analyze the yield.
Linear Regression model for crop yield prediction:
To develop the Linear Regression models for crop
yield prediction, Linear Regression analysis is majorly used
for prediction functions because it provides predictedentity
as a function of depended entities.
Steps for crop yield prediction using Linear
Regression algorithm as follows: Inputs are given in
experimental information set of whether or not data crop
information and soil information and their outcomes
predicted crop yield for the experimental dataset. Some
technique are given
Gather, format and organize the information: only
raw information is scarce to work with the model. The data
should be gathered, soft out as per the requirement and
organization it in such a path, that appropriate results are
obtained. Although redoing, additional vital data can be
included.
 Collect the dataset then preprocessing the dataset
for noise removal process.
 Separate data into testing andtrainingsets:thedata
information must to go partitioned into two sets.
Training set can have greatest rate of the data so as
to train most of the examples to create the yield.
The samples are collected under training set.
Testing set uses the remaining measure of the data
to check however they system is performing.
Fig 3.2 Linear Regression Model for Crop Yield
Prediction
 Apply Linear Regression on trained sets: the model
system depends on upon however complicated the
problem is and also the structurelikewiseshouldbe
selected with the requirement. Though altering,the
development modeling and structure is adjusted.
 Validation the Multiple R square, Adjusted R-
squared and F-Statistic values for this models.
 Apply the trained Linear Regression model on test
set and once again calculate the Multiple R square,
Crop Dataset
Accuracy
Classification
Linear Regression
Validate Coefficient determinate
Multiple R squared
Adjusted R-squared
F-Statistic
Preprocessing
Noise Removal
Partition
Training set
Test set
Check if Value
High or low
Low or high

Adjusted R-squared and F-Statistic values.Compare
the values with completely different models of
Linear Regression models. The model that provides
to be the best model for crop yield prediction.
4. RESULT AND DISCUSSION
4.1 PERFORMANCE METRICS
Within recent analysis, effectiveness related
to carries with it technique reviewed creating use of
Espresso at the side of R-Tool 2.35 are widely used to
implement with this carries with it methodology. The actual
effectiveness carries with it technique through – about once
it involves preciseness, recognition moreover as accuracy.
The final results, means category effectiveness could be
superior through the utilization of LR-PROPOSED being an
optimization technique within the category method.
Precision Value: Precision is calculated as the number of
true positive predictions divided by the total number of
positive predictions.
Recall value: Recall value is specified to as the relevant
datasets that are related to the other request Search.
F measure: F measure test’s accurateness and is define as
the weighted harmonic mean of the precision and recall of
the analysis
AnalyzetocomparisonbetweenK-NearestNeighbor
(Existing System)and Linear Regression (Proposed System)
with parameter evaluation.
Table 4.1.1 Comparisons of Parameter Values
Algorithm/
Parameter
Precision Recall F-
Measure
K-NN 0.84 0.85 0.84
LR-Proposed 0.86 0.87 0.89
Fig.4.1.1 Graphical Parameters Comparison for
Existing K-NN with Proposed LR-Proposed algorithm.
Accuracy: Accuracy gives the required relateddatasetsused
for classification. Compute the proportion of true positive
and true negative in all calculated cases.
Analyzed crop accuracy compared with K-NN and LR-
Proposed
Table 4.1.2 Crop Accuracy Comparison
Algorithm Cotton Sugarcane Turmeric
K-NN 87 84 85
LR-
Proposed
95 96 95
Fig 4.1.2 Crop Accuracy Compared with K-NN and LR-
Proposed.
The efficient classification Linear Regression
algorithm is used to develop the model. This algorithm is
compared, and accuracy is evaluated. From the above table
5.3.2, it is observed that Linear Regression had the best
predictive power with high accuracy as compared to K-
Nearest Neighbor.

CONCLUSION
In accurate prediction of different specified crop
yields across different districts will help to farmers of India.
Yield estimation models are utilized in
preciseness Agriculture to extend yield productiontosatisfy
demand and to recommend to the government in regard to
prediction crop yield on imports of Trichy, Tamilnadu
dataset to avoid overlapping. During this work the
regression approach were tested in their yield prediction
capabilities. The readingswereusedformodel inputs. Linear
regression algorithms offered acceptable estimation
accuracy, though higher prognostic power could also be
obtained by parameters like year, crop, area, production (in
tons) and alternative variables, like climate, agricultural
practices and soil characteristics are including within the
model development. The model using linear regression can
be suggested for Ecuadorian conditions. In yield prognostic
models are not existent for any crop. From this proposed
system the yield of crop (sugarcane, cotton, and turmeric)
are predicted in highest level. This model may be
reformulated using alternative crop assessments within the
future, to develop methods for increasing yield and land
territorial management in alternative crops of importance,
like wheat, rice.
REFERENCES
1) D Ramesh, B Vishnu Vardhan, “Analysis of Crop
Yield Prediction using Data Mining Techniques”,
International Journal of Research in Engineering
and Technology (IJRET),Vol.4, 2015.
2) DakshayiniPatil, Dr. M .S Shirdhonkar, “Rice Crop
Yield Prediction using Data Mining Techniques: An
Overview”, International Journal of Advanced
Research in Computer Science and Software
Engineering, Volume 7, Issue 5, ISSN: 2277 128X
,2017.
3) Dr. Rakesh Poonia1 , Sonia Bhargava “Prediction of
Crops Methodology using Data MiningTechniques”,
International Journal of Innovative Research in
Science, Engineering and Technology, Vol. 6, Issue
10, October 2017.
4) Leemans V, M F Destain, "A Real Time Grading
Method of Apples Based onFeaturesExtractedfrom
Defects", J. Jood Eng., 2004, pages: 83-89.
5) Mehta D R, Kalola A D, Saradava D A, Yusufzai A S,
"Rainfall Variability Analysis andItsImpactonCrop
Productivity - A Case Study", Indian Journal of
Agricultural Research, Volume 36, Issue 1, 2002,
pages : 29-33.
6) Prof .M.S.PrasadBabu, N.V.Ramana Murty,
S.V.N.L.Narayana, “A Web Based Tomato Crop
Expert Information System Based on Artificial
Intelligence and Machine Learning Algorithms”,
IJCSIT, Vol. 1 (1), 2010, 6-15.
7) R J Brooks, “Simplifying Sirius: Sensitivity Analysis
and Development of a Meta-Model for Wheat Yield
Prediction”, European Journal of Agronomy,vol.14,
2001, pages: 43-60.
8) Rajshekhar Borate., “Applying Data Mining
Techniques to Predict Annual Yield of Major Crops
and Recommend Planting Different Crops in
Different Districts in India”, International Journal of
Novel Research in Computer Science and Software
Engineering,Vol. 3, Issue 1, pp: (34-37), April 2016.
9) Ramesh A. Medar and Vijay. S. Rajpurohit “A Survey
of data mining techniques for cropyield prediction”,
IJARCSMS, Volume 2, Issue 9, September 2014 pg.
59-64.
10) Veenadhari, S., Bharat Misra, D Singh, “Data mining
Techniques for Predicting Crop Productivity – A
review article”, IJCST, International Journal of
Computer Science and technology march 2011.

IRJET- Analysis of Crop Yield Prediction using Data Mining Technique to Predict Annual Yield of Major Crops

More Related Content

What's hot (17)

Similar to IRJET- Analysis of Crop Yield Prediction using Data Mining Technique to Predict Annual Yield of Major Crops (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Analysis of Crop Yield Prediction using Data Mining Technique to Predict Annual Yield of Major Crops