Crop Management System Using Machine Learning
Crop Management System Using Machine Learning
Volume 6, Issue 7, July – 2021 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Crop Management System Using
Machine Learning
Sai Anand P, Sandeep S, V Pranav Srinivas Rajeshwari G.L
Computer Science and Engineering Assistant Professor
SJB Institute of Technology Computer Science and Engineering
Bengaluru, India. SJB Institute of Technology
Bengaluru, India.
Abstract:- In India, Agriculture is very important and create new opportunities for discovering, quantifying and
practised broadly across the country. This industry plays comprising data-intensive agricultural processes.
a major economic role in the country's development. So,
three main criteria are needed to be considered while It is also important for our farmers to make profits
growing crops. Choosing fertilizers, selecting proper using scientific approaches and technologies in agriculture.
crops according to the region’s climatic condition and In India, wheat and paddy are the most important food
Knowledge of crop price are the criteria that benefit the crops. Every year farmers in India take difficult decisions
farmers and also helps the economic growth of the that affect their incomes. Using new techniques and digital
country. This paper proposes three different models for applications of machine learning farmers can predict the
each of the above-mentioned criteria. The first model is crop's price, can know the fertilizer for the soil, and plan the
the Fertilizer predictor, which predicts the suitable best crop for that season. Fertilizers not only enriches the
fertilizer that can be used for the given soil constituents. soil health but also it increases holding capacity of water.
The second model is the crop predictor, which predicts So, knowing proper usage of fertilizer helps in getting more
the three most likely growable crops based on the given yield. Harvesting crops in places where climatic conditions
climatic conditions. The third model is the future crop are not suitable for the growth of crop leads to a decline in
price predictor, which predicts the crop price in future yield. Hence proper crop should be selected to grow
based on past trends and prices. according to the conditions. And also having an idea of
trends and price of the crops in upcoming years will result in
Keywords:- Crop Management, Machine Learning, Data obtaining more profits for the farmers.
Visualization, KNN, ARIMA, Random Forest, Logistic
Regression. In the current situation, many farmers do not have
enough knowledge about how to use technology to grow
I. INTRODUCTION crops and also, they are not aware of the benefits they get
from farming. Also, farming can be increased by
The most important occupation practiced in our understanding the crop's performance in the particular land
country is agriculture. The economy of India would have and different environmental conditions. These are some of
fallen if there was no agriculture. In the national GDP of the challenges faced by the farmers every day. In developing
India, around 15-20% is contributed by the agriculture Machine Learning Models, the following are the main
sector [1]. Agriculture plays a very important role in overall challenges. Training data overfitting. Complexity of the data
development as well as in the economic sector of the pipeline. The precision and interpretability of the model
country. Almost 60% of the land in our country is utilized balance. Collection and storage of large quantities of farm
for agriculture to satisfy the needs of the people in our datasets. This paper proposes three different models which
country. are efficient for crop management. The first model predicts
the fertilizer to be used based on soil nutrients and
In the global economy, agriculture plays an important environmental conditions. The second models predict the
role. With the continued growth of the human population, crops which most probable to grow in the given climatic
pressure will increase on the agriculture system. Agriculture, conditions. The third model predicts the future price of the
now also called digital agriculture, has emerged from the crop from the past prices and trends. Also, proper Data
new scientific fields which employ data intensive Visualization techniques are used all the above proposed
approaches to promote farm productivity while minimising models.
environmental effects [2]. Information gathered in advanced
crop cultivation are based on different types of sensors This project aims to help the farmers to cultivate the
which allow a more precise and quick decisions to be taken, best crop, predict the price of wheat, and also it will help to
enhancing operating environment understanding (interaction do the fertilizer analysis.
between flexible crop, soil, and weather). Machine learning
(ML) has emerged, in combination with large data
technology and high-performance computer technology, to
IJISRT21JUL260 www.ijisrt.com 111
Volume 6, Issue 7, July – 2021 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
II.CROP MANAGEMENT IV. METHODOLOGY
Machine learning technique was used to predict the The first step in each of the proposed model to collect
price of the crop. By recognising the patterns in our training suitable datasets. Open-source datasets have been collected
dataset i.e., the input of the model, the price of the crop can from the website kaggles.com. The next step is to apply
be determined. The parameters used in the dataset are of suitable data visualization techniques for the obtained
previous years and their respective prices along with in- datasets. Data Visualization focuses not only on qualitative
season profit, off-season loss and the respective rise and fall descriptions of data but also on the estimation of data. For
of the prices. Using this, the prices of the crops in the qualitative understanding, Data visualization provides an
coming years can be predicted [3]. Productivity and important suite of tools. Techniques that are used in this
operational efficiency can be improved by designing Crop model are Hist plot, Correlation Data, Heatmap, Pair plot,
Management. Storing accurate records, monitoring the team, Bar plot etc. After Visualization, the Data Pre-processing
keeping tabs on chemical applications or inventory, step is followed. Here Data is transformed for the sake of
scheduling and managing production practice events can all convenience of the Machine Learning model to obtain high
be done effortlessly by Crop Management. Soil is one of the accuracy.
most important factors for getting good quality and good
quantity of yield. The use of the correct fertilizer is vital for The next step is the major step, where to extract
this purpose. Our model uses a dataset containing the features, machine learning algorithms are used. Machine
various contents of the soil as parameters to determine the learning algorithms do not involve any human interaction,
appropriate fertilizer to be used for that particular soil [4]. instead, these algorithms learn from the previous
The knowledge of being aware of what crops to grow in experiences which are given as input to these algorithms.
which ecosystem makes all the difference between a good Generally, the learning tasks are functions where it maps
yield and bad yield. Our model helps the user to make this input to the output functions.
choice. The choice of the crop is determined by recognizing
the patterns in our training dataset which is given as the
input to the model. The parameters in the dataset are
climatic conditions such as rainfall, temperature etc. This
will give an overall prediction of which is best crop for that
ecosystem [5]. Real time data is provided by Crop
Management to help growers make right decisions at the
right time [6]. Farmers can view costs of production and
make decisions easily by using crop management system in
order to increase profitability.
III. MODEL ARCHITECTURE
In the proposed systems we are designing the system
for crop management analysis using Machine Learning
approach. These Machine Learning approaches help in data Fig 2 Methodology used for prediction
visualization as well as developing prediction models.
The Machine Learning Algorithm used in our Models
are:
KNN-Model: KNN is an algorithm based on instance-based
learning. It is often called lazy learning. Till the function is
evaluated, all computation is delayed and the function is just
approximated locally. KNN model is used in the first
proposed model that is for fertilizer predictor.
Random Forest: To solve both regression and classification
tasks Random Forest algorithm can be used. Random Forest
solved the given tasks using Bootstrap and Aggregation.
Bootstrap means using multiple decision trees and
Aggregation is commonly known as bagging. Random
Forest algorithm is used in both fertilizer predictor and crop
predictor models.
Logistic Regression: For binary classification-related
problems, Logistic regression is largely used. This algorithm
is used in crop predictor model.
ARIMA Model: “Auto-Regressive Integrated Moving
Average” is shortly called as ARIMA model. This model is
mainly used for forecasting problems. ARIMA model is
Fig 1 Architecture for the Proposed Model used in Future crop price predictor model.
IJISRT21JUL260 www.ijisrt.com 112
Volume 6, Issue 7, July – 2021 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
V. CONCLUSION
While there have been similar projects in the past, this
project aims to bring a more refined and accurate outcome.
The goals have been satisfied with the code developed.
The proposed system in data visualization and data
predictions using various prediction models such as the
KNN model, Random Forests, Arima model helps in
analysing the large datasets and predicting with high
accuracy. These systems give a better way of approaching
analytics as well as predictions. These systems save a large
amount of time and have better accuracy.
REFERENCES
[1]. Mythresh A1, Lavanya B2, Meghana BS3, Nisarga B4,
Crop Prediction using Machine Learning, Students of
Dept. Computer Science and Engineering,
International Research Journal of Engineering and
Technology (IRJET), June 2020
[2]. Konstantinos G. Liakos, Patrizia Busato , Dimitrios
Moshou Simon Pearson ID and Dionysis Bochtis,”
Machine Learning in Agriculture: A Review”, Lincoln
Institute for Agri-food Technology (LIAT), University
of Lincoln, Brayford Way, Brayford Pool, Lincoln
LN6 7TS, UK, [email protected], pg4,2018
[3]. Dhanapal, R., AjanRaj, A., Balavinayagapragathish,
S., & Balaji, J. (2021, May). Crop price prediction
using supervised machine learning algorithms.
In Journal of Physics: Conference Series (Vol. 1916,
No. 1, p. 012042). IOP Publishing.
[4]. Amrutha, A., Lekha, R., & Sreedevi, A. (2016,
December). Automatic soil nutrient detection and
fertilizer dispensary system. In 2016 International
Conference on Robotics: Current Trends and Future
Challenges (RCTFC) (pp. 1-5). IEEE.
[5]. Balakrishnan, N., & Muthukumarasamy, G. (2016).
Crop production-ensemble machine learning model for
prediction. International Journal of Computer Science
and Software Engineering, 5(7), 148.
[6]. Rohit R, Vishnu R, Kishore A, Deeban
Chakkarawarthi, “Crop Price Prediction and
Forecasting System using Supervised Machine
Learning Algorithms”, (BE Department of CSE,
JCTCET Coimbatore, Tamil Nadu, India),
International Journal of Advanced Research in
Computer and Communication Engineering
(IJARCCE) Vol 9, Issue 3 March pg27,2020.
[7]. Kaggles.com for the open-source datasets.
