Prediction of Land Suitability for Crop Cultivation Using Classification Techniques

Ganesan, Mariammal; Andavar, Suruliandi; Raj, Raja Soosaimarian Peter

doi:10.1590/1678-4324-2021200483

Acessibilidade / Reportar erro

Brasil

Brazilian Archives of Biology and Technology

Español English

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

Article - Agriculture, Agribusiness and Biotechnology • Braz. arch. biol. technol. 64 • 2021 • https://doi.org/10.1590/1678-4324-2021200483 copy

Prediction of Land Suitability for Crop Cultivation Using Classification Techniques

Authorship SCIMAGO INSTITUTIONS RANKINGS

HIGHLIGHTS

Comparative analysis of Machine Learning techniques for Crop prediction.

Performance evaluation based on only soil, only environmental characteristics and both.

Performance analysis for classifiers using k-fold validation.

Performance analysis using data splitting method.

Abstract

Agriculture, the backbone of every country, has been an emerging field of research, particularly in the recent past. The soil type and environment are critical factors that drive agriculture, especially in terms of crop prediction. To determine which crops grow best in certain types of soil and environment, the characteristics of the latter are to be ascertained. In the past, farmers picked suitable crops for cultivation, based on first-hand experience. Today, however, identifying appropriate crops for particular areas has become a difficult proposition. The application of machine learning techniques to agriculture is an emerging field of research that helps predicts crops for easy cultivation and improved productivity. In this work, a comparative analysis is undertaken using several classifiers like the k-Nearest Neighbor (kNN), Naïve Bayes (NB), Decision Tree (DT), Support Vector Machines (SVM), Random Forests (RF) and Bagging to help suggest the most suitable cultivable crop(s), based on soil and environmental characteristics, for a specific piece of land. The algorithms are trained with the training data and subsequently tested with the soil and climate-based test dataset. The results of all the approaches are evaluated to identify the best classification techniques. Experimental results show that the bagging method outclasses others with respect to all performance metrics.

Keywords:
agriculture; soil; environmental; crop; machine learning; classification

S.No.	Attributes	Type	Description
S.No.	Soil characteristics
1.	pH (potential of Hydrogen)	Numeric	pH is the main factor for farming
2.	EC (Electrical Conductivity)	Numeric	That affect crop productivity if EC is 0.01 that soil is good soil
3.	OC (Organic Carbon)	Numeric	OC enters the soil through the decomposition of plant and animal residues, root exudates, living and dead microorganisms, and soil biota.
4.	N (Nitrogen)	Numeric	Nitrogen is a key element in plant growth.
5.	P (Phosphorus)	Numeric	Phosphorus helps transfer energy from sunlight to plants, stimulates early root and plant growth, and hastens maturity.
6.	K(Potassium)	Numeric	Potassium increases vigour and disease resistance of plants, helps form and move starches, sugars and oils in plants, and can improve fruit quality.
7.	S (Sulphur)	Numeric	Sulphur is a constituent of amino acids in plant proteins and is involved in energy-producing processes in plants.
8.	Z (Zinc)	Numeric	Zinc helps in the production of a plant hormone responsible for stem elongation and leaf expansion.
9.	B (Boron)	Numeric	Boron helps with the formation of cell walls in rapidly growing tissue. Deficiency reduces the uptake of calcium and inhibits the plant’s ability to use it.
10.	Fe (Iron)	Numeric	Iron is a constituent of many compounds that regulate and promote growth
11.	Mn (Manganese)	Numeric	Manganese helps with photosynthesis.
12.	Cu (Copper)	Numeric	Copper is an essential constituent of enzymes in plants
	Environmental Factors
13.	Texture	Integer	It has major influence on crop growth. It influences aeration, water movement etc.
14.	Season	Integer	Season is the challenging factor for crop growth.
15.	Rainfall	Numeric	Rainfall has the great impact on crop growth. Excessive and insufficient rainfall affects the yield.
16.	Average Temperature	Integer	It is important for growth and development

S. No	Metric	Formula	Range	Description
1.	Accuracy	$\frac{T P + T N}{T P + T N + F P + F N}$	From 0 to 1. Value close to 1 shows the better prediction	The accuracy rate has historically been the most commonly used statistical indicator.
2.	Kappa	$\frac{P_{a g r e e} - P_{c h a n c e}}{1 - P_{c h a n c e}}$	-1 to 1. 1 reflects classification is significantly better than random; less than 1 reflects no better than random.	It is a measure of agreement between two individuals.
3.	Precision	$\frac{T P}{T P + F P}$	0 to 1. Value nearby 1 denotes has less false positive prediction.	The number of true positive predictions divided by the total number of positive predictions is determined as precision.
4.	Recall	$\frac{T P}{T P + F N}$	0 to 1. Value close by 1 means has less false negative prediction	Analyzes of recall can help to examine the main sources of uncertainty of model prediction..
5.	Specificity	$\frac{T N}{T N + F P}$	0 to 1. Value close to 1 indicates has less negative prediction	The proportion of negative results out of the number of samples which were actually negative.
6.	F1 Score	2 * $\frac{P P V . T P R}{P P V + T P R}$	0 to 1. Value close 1 implies has better precision and recall; Value 0 means worst precision and recall	The accuracy of the measurement test and is known as the weighted harmonic mean of the tests ' precision and recall.

Classifiers	Performance Metrics (%)
Classifiers	Accuracy	Kappa	Sensitivity	Specificity	Precision	F1 Score
kNN	0.5625	0.4939	0.6763	0.9438	0.6797	0.6783
NB	0.6812	0.6243	0.7494	0.9577	0.7467	0.7481
DT	0.7125	0.6637	0.7935	0.9622	0.7573	0.7643
SVM	0.7750	0.7362	0.8242	0.9706	0.7853	0.8042
RF	0.8375	0.8092	0.8434	0.9789	0.8420	0.8427
Bagging	0.8875	0.8686	0.9108	0.9790	0.8597	0.8845

Classifiers	Performance Metrics (%)
Classifiers	Accuracy	Kappa	Sensitivity	Specificity	Precision	F1 Score
kNN	0.3500	0.2151	0.8006	0.8804	0.3673	0.5036
NB	0.3512	0.2457	0.8198	0.9017	0.3875	0.5263
DT	0.4435	0.3089	0.8811	0.9154	0.4084	0.5581
SVM	0.4500	0.3433	0.8958	0.9166	0.4516	0.6005
RF	0.4625	0.3612	0.9138	0.9208	0.5017	0.6487
Bagging	0.5400	0.3835	0.9176	0.9266	0.5184	0.6662

Classifier	Performance Metrics (%)
Classifier	Accuracy	Kappa	Sensitivity	Specificity	Precision	F1 Score
kNN	0.6	0.5804	0.7023	0.9473	0.6909	0.6965
NB	0.7854	0.7749	0.8636	0.9584	0.8261	0.8444
DT	0.8188	0.8088	0.8687	0.9647	0.8528	0.8607
SVM	0.8312	0.8456	0.8937	0.9712	0.8743	0.8838
RF	0.8875	0.8681	0.919	0.9796	0.9133	0.9161
Bagging	0.9062	0.8901	0.9257	0.9878	0.9255	0.9256

Classifier	Folds	Performance Metrics (%)
Classifier	Folds	Accuracy	Kappa	Precision	Recall	Sensitivity	F1 Score
Bagging	10	90.62	89.01	92.55	92.57	98.78	92.56
	20	89.3	87.69	91.23	91.25	97.46	91.24
	30	87	85.38	88.92	88.94	95.15	88.93
	40	88.32	86.4	89.94	89.96	96.17	89.95
	50	87.8	85.88	89.42	89.44	95.65	89.43
	60	86.5	84.58	88.12	88.14	94.35	88.13
	70	87.2	85.18	88.72	88.74	94.95	88.73
	80	87.77	85.75	89.29	89.31	95.52	89.3
	90	86.3	85.42	88.96	88.98	95.19	88.97
RF	10	88.65	86.81	91.33	91.9	97.96	91.61
	20	87.93	85.49	90.01	90.58	96.64	90.29
	30	85.63	83.18	87.7	88.27	94.33	87.98
	40	86.95	84.2	88.72	89.29	95.35	89
	50	86.43	83.68	88.2	88.77	94.83	88.48
	60	85.13	82.38	86.9	87.47	93.53	87.18
	70	85.83	82.98	87.5	88.07	94.13	87.78
	80	86.4	83.55	88.07	88.64	94.7	88.35
	90	84.93	83.22	87.74	88.31	94.37	88.02
SVM	10	83.12	84.56	87.43	89.37	97.12	88.38
	20	81.8	83.24	86.11	88.05	95.8	87.06
	30	79.5	80.93	83.8	85.74	93.49	84.75
	40	80.82	81.95	84.82	86.76	94.51	85.77
	50	80.3	81.43	84.3	86.24	93.99	85.25
	60	79	80.13	83	84.94	92.69	83.95
	70	79.7	80.73	83.6	85.54	93.29	84.55
	80	80.27	81.3	84.17	86.11	93.86	85.12
	90	78.8	80.97	83.84	85.78	93.53	84.79
DT	10	81.88	80.88	85.28	86.87	96.47	86.06
	20	80.56	79.56	83.96	85.55	95.15	84.74
	30	78.26	77.25	81.65	83.24	92.84	82.43
	40	79.58	78.27	82.67	84.26	93.86	83.45
	50	79.06	77.75	82.15	83.74	93.34	82.93
	60	77.76	76.45	80.85	82.44	92.04	81.63
	70	78.46	77.05	81.45	83.04	92.64	82.23
	80	79.03	77.62	82.02	83.61	93.21	82.8
	90	77.56	77.29	81.69	83.28	92.88	82.47
NB	10	78.54	77.49	82.61	86.36	95.84	84.44
	20	77.22	76.17	81.29	85.04	94.52	83.12
	30	74.92	73.86	78.98	82.73	92.21	80.81
	40	76.24	74.88	80	83.75	93.23	81.83
	50	75.72	74.36	79.48	83.23	92.71	81.31
	60	74.42	73.06	78.18	81.93	91.41	80.01
	70	75.12	73.66	78.78	82.53	92.01	80.61
	80	75.69	74.23	79.35	83.1	92.58	81.18
	90	74.22	73.9	79.02	82.77	92.25	80.85
kNN	10	60	58.04	69.09	70.23	94.73	69.65
	20	58.68	56.75	67.77	68.91	93.41	68.33
	30	56.38	54.44	65.46	66.6	91.1	66.02
	40	57.7	55.46	66.48	67.62	92.12	67.04
	50	57.18	54.94	65.96	67.1	91.6	66.52
	60	55.88	53.64	64.66	65.8	90.3	65.22
	70	56.58	54.24	65.26	66.4	90.9	65.82
	80	57.15	54.81	65.83	66.97	91.47	66.39
	90	55.68	54.48	65.5	66.64	91.14	66.06

Instituto de Tecnologia do Paraná - Tecpar Rua Prof. Algacyr Munhoz Mader, 3775 - CIC, 81350-010 Curitiba PR Brazil, Tel.: +55 41 3316-3052/3054, Fax: +55 41 3346-2872 - Curitiba - PR - Brazil
E-mail: babt@tecpar.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] *Correspondence: suba.g1212@gmail.com; Tel.: +91-7200668118 (M.G.).