Soil Testing And Classification For The Purpose Of Cultivation Using

M.L

Harsh Ahuja, Vaidehi Gade, Suchi Kumari and Sagar

Department of Computer Science and Engineering
MIT School of Engineering
MIT Art, Design and Technology University
Loni Kalbhor, Pune 412201, Maharashtra, India
{harshkahuja, Vaidehig99, Ksuchi914, Sagar.rishi25}@gmail.com

Prof. Swapnil Patil

Department of Computer Science and Engineering
MIT School of Engineering
MIT Art, Design and Technology University
Loni Kalbhor, Pune 412201, Maharashtra, India
swapnil.patil@mituniversity.edu.in

Abstract— In a country like India where agriculture is the
backbone of our economy, it is very important to identify
the soil types and decide which type of crops to grow.
The objective of our model is to provide a solution for
intelligent agriculture by classifying soil type which can assist
the farmers in increasing productivity to a great extent.
The proposed system applies machine learning to identify
the pattern among data and then process it as per input
conditions. We will use physical properties like texture,
density etc. by doing ribbon testswe, squeeze tests and jar
tests to train the model. We will be developing a user-friendly
GUI using the selected machine-learning technique. It will
be very cost-efficient for the farmers.
Index Terms— Physical Features, Crop Prediction, Soil Clas-
sification, Machine Learning.
used to evaluate the model and compute the results. The
data sets are divided into two categories, the training data
set and the test data set, to create the model. Various
machine learning algorithms are then applied to classify
soils as clay, loamy, or sandy using the soil’s physical
properties. Crops are also recommended based on the
predicted soil type. Algorithms evaluated include SVM,
Decision Trees, Naive Bayes, and Linear Regression, all
available for machine learning models.
This paper is structured as follows: Section II provides
details on the literature survey. Section III explains differ-
ent algorithms used for soil type classification and crop
recommendation; followed by conclusions and references.

I. INTRODUCTION II. LITERATURE SURVEY

Agriculture is the art of cultivating soil, growing crops,
and raising livestock. It is based on the preparation of
plant and animal products for human use and their
distribution on the market. Agriculture provides most
of the domestic food and textiles. Agriculture plays a
very important role in the global economy. The world’s
population is growing very rapidly, and as the population
grows so does the need for food. Farmers’ traditional
methods cannot meet the sheer volume of demand and
hence identifying the right type of soil becomes very
important.
Machine learning in agriculture is used to improve the
product quality of crops in the agricultural sector. Machine
learning is the scientific field that gives machines the
ability to learn without human intervention. The main
objective of this study is to design soil type prediction and
crop recommendation models using machine learning
supervised models. Soil and plant datasets are collected
from his website for online datasets. These data sets are
Here we cover several papers related to soil analysis
and crop yield prediction using various machine learning
techniques and some of their confrontations. Machine
learning is decision-making to solve problems that need
to be modeled from data.
1. Classification of Soil and Crop Suggestion using
Machine Learning Techniques- This model predicts soil
type and suggests a suitable crop that can be cultivated
in that soil. but it has some limitations: it does not contain
real time data.
2. Soil Classification and Crop Suggestion using Ma-
chine Learning Algorithm- This ML model contains data
of different parameters affecting soil like temp, humidity,
pH.These parameters are used to find best suited crops
from all crops, but the model is time consuming and more
costly for home gardeners and farmers.
3. Soil Classification and Crop Suggestion using Ma-
chine Learning- This model takes the data-sets on the
basis of chemical, geographical and climatic properties

1
and predicts the crop based on these parameters.but a D. Support Vector Machine (SVM)
large amount of data-set is required. Support vector machines, or SVMs, are one of the
4. Soil analysis and crop fertility prediction using ma- most popular supervised learning algorithms used for
chine learning- This model is proposed for predicting the both classification and regression problems. The goal of
soil fertility and crop yield with types of the crop that can the SVM algorithm is to create optimal lines or decision
grow on fertile soil. The research is done on soil data-sets boundaries that can divide the n-dimensional space into
and crop data-sets of a specific region. classes so that new data points can be easily placed in
the correct category in the future. This optimal decision
III. METHODOLOGY
boundary is called a hyperplane.
A. Dataset
There are two different datasets are used for the pre-
diction model.
I Soil Dataset
II Crop Dataset
I Soil Dataset: It consists of attributes like pH, rainfall,
humidity, temperature, squeeze test, ribbon test. From
these all attributes we classified and analyzed by applying
ML model which type of soil is this.
II Crop Dataset: It consists of rainfall, humidity, pH, soil
type. From all these attributes we recommended suitable
crops to be growth in that particular soil type.
Fig. 2. Support Vector Machine
B. Method and Experimentation
Datasets are collected from various resources and the
data is categorized and grouped into two sets: E. Decision Tree
I Training Dataset A decision tree is a flowchart-like tree structure where
II Testing Dataset each inner node specifies a test for an attribute, each
branch represents the result of the test, and each leaf
C. Implementation using ML Algorithm node (terminal node) contains a class label. Trees can be
When visiting our website, users are directed to ques- "learned" by "recursive splitting".
tionnaire. Based on the questionnaire responses, these
responses are provided as input to the ML soil type predic-
tion model, which predicts soil type. After estimating the
soil type, we recommend crops. Below is a basic flowchart
of the ML model.

Fig. 3. Decision Tree

F. Logistic Regression
It helps predict the likelihood of an event occurring or
a decision being made. In this analytical approach, the
dependent variable can be finite or categorical, A or B
(binomial regression) or a set of finite options A, B, C, or
D (multinomial regression). Used in statistical software to
understand the relationship between a dependent variable
and one or more independent variables by estimating
Fig. 1. ML Model Flowchart probabilities using the logistic regression equation.
.

2
 made us understand the value of working together in a
group. Finally, we would like to thank each and every
faculty member, co-students who helped us directly or
indirectly to make our idea into a working project.

Fig. 4. Logistic Regression
G. Naive Bayes Classifier
A naive Bayes classifier is a classification method that
evolves from Bayes’ theorem to maximize posterior prob-
abilities and can also be used to increase the classification
probabilities of a variable and the conditional factor range
of parameters used in the classification process. . The class
variables required to determine classification without the
entire covariance matrix.
R EFERENCES
[1] Classification of Soil and Crop Suggestion using Machine Learning
Techniques [online]. Available: https://www.ijert.org/
classification-of-soil-and-crop-suggestion-using-mach
ine-learning-techniques.
[2] SOIL CLASSIFICATION AND CROP SUGGESTION USING MA-
CHINE LEARNING [online]. Available: https://www.irjet.net/
archives/V7/i6/IRJET-V7I6691.pdf
[3] Soil Analysis and Crop Fertility Prediction using Machine Learning
[online]. Available: https://www.academia.edu/46014132/
Soil_Analysis_and_Crop_Fertility_Prediction_using_
Machine_Learning
[4] A comprehensive review on soil classification using deep
learning and computer vision techniques [online]. Available:
https://www.researchgate.net/publication/348913407_
A_comprehensive_review_on_soil_classification_using_
deep_learning_and_computer_vision_techniques

Fig. 5. Naive Bayes

IV. CONCLUSIONS
In this paper, a model is proposed for predicting the
soil type and the crops that can be grown in this soil. The
research has been done on soil datasets and crop datasets
of the Indian region. The model has been tested by using
a different machine learning algorithms.

ACKNOWLEDGMENT
We owe a debt of sincere gratitude, and respect to our
guide and mentor Prof. Swapnil Patil, MIT Art, Design
and Technology University, Pune for his sagacious guid-
ance, vigilant supervision and valuable critical apprecia-
tion throughout this project work. We would also like to
thank our principal Dr. Kishore Ravande, honorable HOD
ma’am and our project coordinator for giving us such an
auspicious opportunity to convert our idea into a working
project. We came to know about a lot of different new
things both theoretical and practical by working on this
major-project. It not only imparted knowledge but also