TENSYMP2022 1570789058

A Review on Prediction of Crop Yield using

1  
2  
Machine Learning Techniques
3  
Mr. Sunil G L Dr. Nagaveni V Mrs. Shruthi U
4   Saividya Institute of Technology Acharya Institute of Technology Presidency University
5   Bengaluru, India Bengaluru, India. Bengaluru, India
6   [email protected] [email protected] [email protected]
7  
8  
9  
10   Abstract— Population is increasing day by day in India 2. Feature Selection: A subset of attributes are selected
11   and demand for food is also increasing, due to this reason, from the crop knowledge base to create a data set with
12   agriculture is very essential. Advanced technologies like data different parameters like area, yield, climate, soil attributes,
13   mining techniques, machine learning techniques, remote fertilizers, etc.
14   sensing and image processing, etc., can be used for prediction
of crop yield. Machine Learning techniques play a significant 3. Yield Prediction: crop yield can be predicted by
15   role in crop yield prediction using different parameters like applying Machine learning supervised or unsupervised
16   temperature, rain fall, soil parameters, area of sowing, etc,. algorithms.
17   This paper carried out a comparative study on various
18   machine learning techniques like supervised and unsupervised 4. Pattern Visualization and Prediction Rules: Then
19   learning methods to predict crop yield. It has been observed prediction rules will be applied to the predicted result and the
patterns are visualized.
20   that classification techniques give high accuracy in prediction
21   of multiple crop yields with different parameters than 5. Evaluation: The performance is evaluated between
regression and unsupervised learning methods
22   the provided samples in the crop knowledge base and
23   Keywords— Crop yield prediction, Supervised Learning;
predicted results. The farmer or expert’s interaction may
24   require to find out the efficiency predicted results using
2022 IEEE Region 10 Symposium (TENSYMP) | 978-1-6654-6658-5/22/$31.00 ©2022 IEEE | DOI: 10.1109/TENSYMP54529.2022.9864482

Classification; Regression; Unsupervised learning techniques.

25   machine learning techniques. The general block diagram of
26   I. INTRODUCTION the crop yield prediction system is shown in Figure 1.
27   Agriculture is an important source of economic
28   development. To improve the production in agriculture solve
29   the demand for food, marketing problems [1] inadequate use
30   of form equipment, plant diseases detection, [2][3] and crop
31   yield prediction[4], etc. The Crop yield mainly depends on
32   weather conditions. Today youngsters are very much
33   interested in the farming system because the current status in
34   Indian agriculture is very satisfactory and they are using
advanced technologies to make this agriculture more
35  
prestigious.
36  
37   Crop yield prediction is an important factor in food
38   security assessment and policymaking. Crop growth and
39   yield data monitoring is a tedious task in cultivation. Farmers
40   are more concerned about yield prediction based on climate
41   parameters, soil parameters, fertilizers, diseases. Most of the
42   research works in agriculture using a biological mechanism
to recognize the growth of a crop and improvements in yield.
43  
Nowadays, because of the availability of enormous data in
44   agriculture, many research works are going on using
45   different techniques for yield prediction.
46  
47   Machine learning and Data Mining are most extensively
48   used technologies in agricultural problems. These methods
49   used for extracting important and useful information from a Fig. 1. Crop Yield Prediction Architecture.
large dataset. To predict the crop yield, data mining process
50  
can be used by extracting the feature information from a data III. LITERATURE SURVEY
51  
set and convert into an intelligible structure. The following
52   sections contain a general architecture for crop yield An extensive literature survey has been conducted to
53   prediction and a literature review on supervised and identify machine learning models for the crop yield
54   unsupervised machine learning techniques. prediction. We studied different supervised and unsupervised
55   learning algorithms developed for prediction of crop yield. A
56   II. GENERAL CROP YIELD PREDICTION SYSTEM summarized review carried out on the collected publications
57   is given in Table. 1 based on our knowledge. The following
1. Input: Collect data from formers, agricultural sub-sections discussed different classification and regression
60   department, and meteorological department with different supervised learning models and clustering and association
61   parameters.
62  
63  
64  
65  
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on August 31,2024 at 12:04:54 UTC from IEEE Xplore. Restrictions apply.
1
rule-based unsupervised learning models developed for yield regression techniques In analysis Boosted Random Tree
prediction. method got more than 87% in all years and Random Forest
Regression is a stable method to predict the maize yield. [20]
A. Supervised Learning Compared MLR, ANN and proposed new model hybrid
In supervised machine learning methods, predictions can MLR-ANN to predict paddy crop yield and Hybrid MLR-
be done by the machines which are trained using the training ANN provides better result in RMSE, MAE value.
data consisting of labels with the correct output. Major
supervised learning methods used in crop yield prediction are [21] MLR, Fuzzy Logic, ANFIS Models model
classification and regression techniques. developed for the wheat crop. Based on the RMSE value
ANFIS Technique is better than Fuzzy logic and MLR
Classification: Prediction of the yield of a new crop can techniques. [4] Researchers presented Multi Linear
be done using the classification model, where the Regression technique and Density Based Clustering for
classification model groups the yield of crops based on their multiple crop prediction and compare the results. Multi
previous record or information regarding the same crop. [5] Linear Regression techniques provide better results
Authors used different classification techniques are K-NN, compared to Density Based Clustering.
LR and ANN for prediction of multiple crop yield. In
experimental results ANN Techniques provides a better [22] Fuzzy time series Regression used for prediction of
prediction for Rice, Potato, Wheat Crop and Linear Rice crop and proved is robust. [23] Stacked regression,
Regression provides better prediction- Boro, Amon Crop. Kernel Ridge, Lasso, Enet advanced regression algorithm
developed for multi crops and in result stacked regression got
[6][7][8] Weka tool used for Prediction of crops using RMSE value less than 1 compared to other techniques.
different classifier algorithms like J48, LWL, LAD Tree,
IBK. J48 & LAD Tree achieved the highest accuracy, B. Unsupervised Learning Algorithm
sensitivity, and specificity for rice crop and IBK achieves In unsupervised machine learning methods, machines are
Highest Accuracy for multiple crops. trained using unlabelled training data to find hidden patterns
and insights. Major unsupervised learning methods used in
[9] Multilayer Perceptron Neural Network for rice crop crop yield prediction are Clustering and association rule. [24]
gives the accuracy 97.5% with a sensitivity of 96.3% and Association rule technique used for the wheat crop to find
specificity of 98.1%. [10] Bayes Net and Naïve Bayes the relationship between rain fall, Crop variety & soil and
technique for rice crop yield developed and the performance concluded that the Association rule can be used to provide
of Bayes Net was better than Naive Bayes. valuable information for crop yield. [25] Fuzzy C Mean
[4] Prediction of Rice crop yield developed using SVM Clustering Algorithm used for prediction of wheat crop with
method and the experimental result shows better when neural network and got RMSE Value 0.17. [26] K-Mean,
compared with Naïve Bayes, Bayes Net, and MLP. [11] The DBSCAN, CLARA, PAM clustering techniques developed
proposed algorithm for prediction for wheat crop remote for soil Classification for Bangladesh, and in Results PAM
sensing data yield provides better result than PCA. [12] The come out best. For yield prediction Multiple Linear
researcher analyzed Rapeseed Crop Prediction using ANN Regression, Simple Linear Regression, Stepwise Linear
with MLP and in the Mean Absolute Percentage error less Regression, Generalized Linear Models are used. In result
than 9.43%. analysis, Generalized Linear Model shows the Lowest
RMSE Value.
[13] Random Forest and Gradient Boosting Tree model
developed for sugar cane and in experimental result random Machine Learning algorithms comparisons used in crop
forest obtained better result than Gradient Boosting Tree. yield prediction is summarized and presented in Table 1.
[14] To predict rice crop yield authors used Recurrent Neural
Network using soil parameters, Rainfall, yield. In analysis, TABLE I. COMPARISON OF MACHINE LEARNING
the Root Mean Square Value is less than 45% in all years. TECHNIQUES.
[15] RNN-CNN Framework, Random forest, Deep fully Techniques Culture Parameters Result
connected neural network algorithms are developed to
predict soyabean and corn crop yield and they got better
results in RNN-CNN framework with RMSE value 4.15. Classification
[16] Deep Recurrent Q-Network model proposed for
predicting paddy crop and got an accuracy of 93.7%. LR, K-NN, Rice, Rainfall, ANN
ANN [5] Potato, Humidity, provides a
Regression: Regression models are used to analyze the
Wheat, Temperature better
environmental factors and their influence on their crop yield.
Many researchers are used both regression and classification Boro, , Average prediction
techniques for prediction of crop yield. [17] An analysis amon Sunshine, for Rice,
using a Multi Linear Regression done for predicting the tea Biotic Input Potato, and
leaf yield with climatic parameters for the past 30 years in Attributes Wheat.
Assam and model got accuracy of 82%. [18] Multi linear LR provides
regression, KNN classifier and ANN model developed to a better
predict yield of ground nut and result found that K-nearest prediction
classifier better result with RMSE value. for Boro and
amon.
[19] Operational Land Imagess from remotely sensed
landsat are used for maize crop yield using different

Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on August 31,2024 at 12:04:54 UTC from IEEE Xplore. Restrictions apply.
2
J48, LAD Rice Rainfall, Result in ANN [12] Rape Temperature Mean
Tree, LWL, Temperature accuracy Seed , Absolute
IBK [6] , Soil J48-97.53%, Atmospheric Percentage
Production LWL- Precipitation error less
And Yield 95.69% than 9.43%
LADTree-
85.66 % Random Sugar Soil, Area, Result
IBK-81.62 % Forest, Cane Groove 71.83%
Gradient Width, accuracy for
J48, LAD Rice Temperature Accuracy Boosting Yield, Random
Tree, LWL, , Soil, Area, J48-97.43%, Tree [13] Resource of Forest
IBK [7] Production, LWL- Water, And 71.64%
Yield 95.38% Method of accuracy for
LAD Tree- Irrigation, Gradient
97.54% Epidemic, Boosting.
IBK-94.93% Fertilizer
And Rainfall
J48, LWL, Multiple Year, Crop, Accuracy
LAD Tree, Crop Area, and IBK- RNN [14] Rice Soil, The error
IBK [8] Production. 80.794% Rainfall rate is less
LAD Tree- than 45%
62.25%
J48- CNN-RNN , Corn and Solar RMSE Value
78.145%, RF-NN [15] soybean Radiation, is 4.15 in
LWL- Min & max CNN-RNN
66.22% Temperature compared to
other
MLP Neural Rice Rainfall, Result methods.
Network [9] Temperature Accuracy-
Area, 97.5% DR-Q- Paddy Temperature Deep
Production, Sensitivity- Network , Recurrent Q-
Yield, 96.3% model, DL, Precipitation Network
Crop Specificity- ANN, RF , Area model
Evapotranspi 98.1% [16 ] Cultivated, provides
ration Paddy 93.7 %
Production, accuracy
Bayes Net, Rice Rainfall, BayesNet- Yield
Bayesian Temperature 97.53% Acquired
network Area, accuracy
[10] Production, Bayesian
Yield, network- Regression
Crop 84.69%
Evapotranspi accuracy MLR [17] Tea Rainfall, Max.
ration Temperature Accuracy-
, Relative 82%,
Support Rice Rainfall, 78.76% Humidity, Min
Vector Temperature accuracy, Evaporation Accuracy-
Machine , Area, 68.17% And 65%
[26] Production sensitivity Sunshine
And Yield and 83.97%
specificity. MLR, K- Ground Rainfall, K-nearest
NN, ANN Nut Min & Max Neighbor
PCA, ANN, wheat Wheat Leaf ANN [18] Temperature classifier
GA [11] Data achieved a , Soil better results
Collected better result. Parameters with a 1.23
Through RMSE value.
Remote
Sensors

Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on August 31,2024 at 12:04:54 UTC from IEEE Xplore. Restrictions apply.
3
BRT, RF Maize Remotely BRT method K-Mean, Soil Medium and PAM comes
SVM, GPR Sensed Land higher than PAM, High Land out best
[19] Set 8 OLI 87% for all CLARA,DB
Images years. SCAN [26]
NDVI Data RFR stable
Sets Derived method to
Form predict the IV. CONCLUSIONS
maize yield. Numerous research works have been conducted using
machine learning for crop yield prediction. We carried out a
MLR, ANN, Paddy Irrigation Hybrid comparative study on supervised and unsupervised machine
Hybrid Area, MLR-ANN learning techniques for prediction of crop yield. In the
MLR-ANN Rainfall, provides supervised learning method, many authors developed
[20] Max & Min better results classification and regression techniques for crop yield
Temperature in RMSE, prediction. Comparatively, parameters are more in
, Solar MAE value. classification techniques than regression, and also
Radiation classification techniques achieved better results than
regression techniques in the prediction of multiple crop
FL- ANFIS Wheat Soil Water RMSE yields. In the Unsupervised method technique, clustering and
and (esw), Rain Values association rule mining methods were developed. We found
MLR & Radiation FL-6.4251 that, a lot of scope in research using clustering for yield
[20] ANFIS- prediction. In this emerging and interesting research field we
3.3282 can carry out our research on unsupervised learning methods,
MLR-9.525. Future work can be carried out to predict crop yield using Bi-
Clustering will help solve a complex problem.
MLR, DB Multiple Year, MLR
Cluster Crop Rainfall, provides a
[4] Area Of better result. REFERENCES
Sowing, [1] S Krishna, M S Midul, Rahul R Pillai, “An Efficient
Yield, Agricultural Marketing System for Optimizing the
Benefits of Farmers”, 2nd International Conference on
Fertilizers, Intelligent Computing, Instrumentation and Control
Production Technologies, Kannur, 2019.
[2] U. Shruthi, V. Nagaveni and B. K. Raghavendra, "A
Fuzzy Time Rice Crop Yield the proposed Review on Machine Learning Classification Techniques
[22] Data algorithm is for Plant Disease Detection," in 5th International
Conference on Advanced Computing &
robust Communication Systems (ICACCS), Coimbatore, 2019.
[3] U. Shruthi, V. Nagaveni, C. S. Arvind, and G. L. Sunil,
SR, KR Multi Rainfall, Stacked "Tomato Plant Disease Classification Using Deep
Lasso, Enet Crops Soil Type, regression Learning Architectures: A Review," in Algorithms for
Intelligent Systems. India: Springer, Singapore, 2022,
[23] Nitrogen, results less ch. 15, pp. 153–169.
Potassium than 1 in [4] S. H. A. a. M. H. Aditya Shastry, "A Parameter based
RMSE. ANFIS Model for crop yield prediction," in IEEE
International Advance Computing Conference,
Bangalore, 2015.
Unsupervised Learning Algorithm [5] N. T. M. N. H. M. T. K. A.T.M Shakil Ahamed,
"Applying Data Mining Techniques to Predict Annual
Yield of Major Crops and Recommend Planting
Fuzzy C Wheat Biomass, RMSE 0.17 Different Crops in Different Districts in Bangladesh," in
Mean Crop Temperature SNPD, Takamatsu, Japan, 2015.
Cluster [25] , Rainfall, [6] L. J. A. Niketa Gandhi, "Rice Crop Yield Forecasting
of Tropical Wet and Dry Climate Zone of India Using
And Solar Data Mining Techniues," in 2016 IEEE International
Radiation Conference on Advances in Computer Applications
(ICACA), Coimbatore, India, 2016.
Association Wheat Crop Association [7] L. J. A. S. M. N. Niketa Gandhi, "Applications of data
Rules [24] Variety, rules can be mining techniques for predicting rice crop yield in
Semi-Arid Climate zone of India," in 2017 IEEE
Rainfall, used to Technological Innovations in ICT for Agriculture and
Soil provide Rural Development, Chennai, India , 2017.
valuable [8] P. P. S. C. S. I. Shruti Mishra, "Use of Data Mining in
information Crop Yield Prediction," in Second International
Conference on Inventive Systems and Control (ICISC
for crop 2018), Coimbatore, India, 2018.
yield. [9] O. P. J. A. Niketa Gandhi, "Rice Crop Yield Prediction
Using Artificial Neural Networks," in IEEE
International Conference on Technological Innovations

Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on August 31,2024 at 12:04:54 UTC from IEEE Xplore. Restrictions apply.
4
 in ICT For Agriculture and Rural Development, [26] A. T. K. M. M. R. A. M. R. M. A. R. M. R. Sadia Afrin,
Chennai, India, 2016. "Analysis of Soil Properties and Climatic Data to
[10] L. J. A. P. Niketa Gandhi, "PredictingRice Crop Yield Predict Crop Yields and Cluster Different Agricultural
Using Bayesian Networks," in 2016 Intl. Conference on Regions of Bangladesh," in 17th International
Advances in Computing, Communications and Conference on Computer and Information Science
Informatics (ICACCI),, Jaipur, 2016. (ICIS), Singapore, 2018.
[11] Q. Q. Yao Zhang, "Winter Wheat Yield Estimation [27] N. G. O. P. P. Niketa Gandhi, "Rice Crop Yield
With Ground Based Spectral Information" in 2018 Prediction in India using Support Vector Machines," in
IEEE International Geoscience and Remote Sensing 2016 13th International Joint Conference on Computer
Symposium, Milan, Italy,, 2018. Science and Software Engineering (JCSSE), Khon
Kaen, Thailand, 2016.
[12] G. Niedbała, "Simple model based on artificial neural
network for early prediction and simulation winter rape
seed yield" Journal of Integrative Agriculture, vol. 18,
no. 1, pp. 54-61, 2019.
[13] P. C.-U. Phusanisa Charoen-Ung, "Sugarcane Yield
Grade Prediction using Random Forest and Gradient
Boosting Tree Techniques," in 2018 15th International
Joint Conference on Computer Science and Software
Engineering (JCSSE), Nakhonpathom, Thailand, 2018.
[14] S. Kulkarni, S. N. Mandal, G. S. Sharma, M. R.
Mundada and Meeradevi, "Predictive Analysis to
Improve Crop Yield using a Neural Network Model," in
2018 International Conference on Advances in
Computing, Communications and Informatics
(ICACCI), Bangalore, India, 2018.
[15] L. W. a. S. V. A. Saeed Khaki, "A CNN-RNN
Framework for Yield Prediction," Frontiers in Plant
Science, vol. 10, 2020.
[16] D. E. P. M. D. VINCENT, "Crop Yield Prediction
Using Deep Reinforcement Learning Model for
sustainable Agrarain Applications," IEEE Access, vol.
8, 2020.
[17] R. M. B. R. L. S. Rupanjali D. Baruah, "Use of Data
Mining Technique for Prediction of Tea Yield in the
Face of Climate Change of Assam,India," in
International Conference on Information Technology,
Bhubaneswar,, 2016.
[18] P. S. Vinita Shah, "Groundnut Crop Yield Prediction
Using Machine Learning Techniques," International
Journal of Scientific Research in Computer Science,
Engineering and Information Technology, vol. 3, no. 5,
2018.
[19] M. A. D. A., H. S. S. a. S. R. Hossein Aghighi,
"Machine Learning Regression Techniques for the
Silage Maize Yield Prediction Using Time-Series
Images of Landsat 8 OLI," IEEE Journal Of Selected
Topics In Applied Earth Observations And Remote
Sensing, vol. 11, no. 12, 2018.
[20] P. M. G. R. Bhargavi, "A novel approach for efficient
crop yield prediction," in Computers and electronics in
agriculture, 2019.
[21] S. H. A. a. M. H. Aditya Shastry, "A Parameter based
ANFIS Model for crop yield prediction," in IEEE
International Advance Computing Conference,
Bangalore, 2015.
[22] B. G. B. Anshul Garg A, "A Robust and Novel
Regression Based Fuzzy Time Series Algorithm for
Prediction of Rice Yield," in International Conference
on Intelligent Communication and Computational
Techniques, Jaipur, 2017.
[23] P. S. V. B. L. A. B. J. Potnuru Sai Nishant, "Crop Yield
Prediction based on Indian Agriculture using Machine
Learning Techniues" in International Conference for
Emerging Technology, Belgaum, 2020.
[24] S. A. N. Leisa J. Armstrong, "Agricultural Decision
Support Framework for Visualization and Prediction of
Western Australia Crop Production," in 3rd
International Conference on Computing for Sustainable
Global Development (INDIACom), New Delhi, 2016.
[25] A. J. S. B. Akanksha Verma, "Crop Yield Prediction of
Wheat Using Fuzzy C Means Clustering and Neural
Network," International Journal of Applied Engineering
Research, vol. 13, no. 11, 2018.

Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on August 31,2024 at 12:04:54 UTC from IEEE Xplore. Restrictions apply.
5