A Review On Agriculture Crop Prediction Techniques Using Machine Learning
A Review On Agriculture Crop Prediction Techniques Using Machine Learning
A Review On Agriculture Crop Prediction Techniques Using Machine Learning
https://doi.org/10.22214/ijraset.2021.39170
International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.429
Volume 9 Issue XII Dec 2021- Available at www.ijraset.com
Abstract: Machine learning is a useful decision-making tool for predicting crop yields, as well as for deciding what crops to
plant and what to do during the crop's growth season. To aid agricultural yield prediction studies, a number of machine learning
techniques have been used. We employed a Systematic Literature Review (SLR) to extract and synthesize the algorithms and
features used in crop production prediction research in this investigation This paper provides a comprehensive overview of the
most recent machine learning applications in agriculture, with a focus on pre-harvesting, harvesting, and post-harvesting issues
The papers have been studied in depth, analysed the methodology and features employed, and made recommendations for future
study. Temperature, rainfall, and soil type are the most commonly utilised features, according to our data, while Artificial
Neural Networks are the most commonly employed method in these models.
I. INTRODUCTION
Food is believed to be a basic human requirement that may be met through farming. Agriculture not only meets humans' basic
requirements, but it is also a source of employment around the world. Agriculture is regarded as the economic backbone and a
source of employment in emerging countries such as India. Agriculture accounts for 15.4 percent of India's GDP. A lack of
information at each level of agriculture creates new problems or exacerbates existing ones, increasing the expense of farming. The
demand on the agriculture industry is increasing as the population grows. Overall losses in agriculture systems, from crop selection
through product sale, are extremely substantial. Keeping track of information on crops, the environment, and the market, as the old
phrase goes, can help farmers make better decisions and solve challenges related to agriculture. To collect and process data,
technologies such as blockchain, IoT, machine learning, deep learning, cloud computing, and edge computing can be used.
Computer vision, machine learning, and IoT applications will assist farmers and related domains enhance production, improve
quality, and ultimately increase profitability. Pre-harvesting, harvesting, and post-harvesting are the three major categories of
agricultural activity. Machine learning advancements have aided in boosting agricultural gains. Machine learning is a recent
technology that is assisting farmers in reducing farming losses by offering detailed crop recommendations and insights.. Machine
learning has by far been the most helpful and an innovative method that is recommended to be used in agriculture based
applications. A farmer can receive an accurate estimate of harvestable versus non-harvestable acres on a particular day by logging
into a customised dashboard on a computer or tablet using machine learning technology. Harvestable crop weight and maturity can
also be assessed and projected. Additionally, crops can be examined both before and after harvest for the presence of attractive
traits, level of damage (if applicable), nutritional makeup, and other factors that may effect the ultimate viable yield and product
pricing using a range of technologies, including image analysis. Machine learning (ML) methodologies are applied in a variety of
industries, from evaluating customer behaviour in supermarkets (Ayodele, 2010) to predicting customer phone usage (Witten et al.,
2016). Agriculture has been using machine learning for several years (McQueen et al., 1995). Crop yield prediction is one of
precision agriculture's most difficult problems, and numerous models have been suggested and confirmed so far. Because crop
production is affected by a range of factors such as climate, weather, soil, fertiliser use, and seed variety, this challenge necessitates
the use of many datasets (Xu et al., 2019). This suggests that predicting agricultural yields is not a simple operation; rather, it entails
a series of complex stages. Crop yield prediction programmes can now reasonably forecast actual yields, but greater yield prediction
performance is still desired Researchers have contributed in various ways to solving the problem of climate change's impact on crop
productivity. In this section, we give a brief literature review of colleague research scholars' work and attempt to critically evaluate
it so that we can determine the scope for future research. as part of our research, we will continue to work.
Its applications can be divided into four major categories:
1) Crop management
2) Livestock management
3) Water management
4) Soil management
A. Pre-processing of Data
The method for acquiring data depends on the sort of project we want to create. For example, if we want to create an ML project
that uses real-time data, we can create an IoT system that uses data from various sensors. The data set can come from a variety of
places, including a file, a database, a sensor, and many other places, but it cannot be utilised immediately for analysis since there
may be a lot of missing data, extremely big values, disorganised text data, or noisy data. As a result, Data Preparation is completed
to address this issue.
One of the most crucial processes in machine learning is data pre-processing. It is the most crucial step in improving the accuracy of
machine learning models. There is an 80/20 rule in machine learning. Every data scientist should devote 80% of their effort to data
pre-processing and 20% of their time to real analysis.
B. Data Classification
When the goal variable is categorical (i.e., the output may be classified into classes – it belongs to Class A, B, or something else), a
classification problem arises.
To begin training a model, we divide it into three sections: 'Training data,' 'Validation data,' and 'Testing data.'You use a 'training dat
a set' to train the classifier, a 'validation set' to modify the parameters, and a 'unseen test data set' to test the classifier's performance.
It'worth noting that only the training and/or validation sets are available to the classifier during training. The test data set must not b
e used when the classifier is being trained. The test set will only be available while the classifier is being tested.
IV. CONCLUSION
According to the findings, depending on the scope of the research and the availability of data, the selected publications employ a
variety of elements. Each paper looks into yield prediction using machine learning, however the features vary. The scale, geological
location, and crop of the research all varies. The features chosen are determined by the dataset's availability and the research's goal.
According to studies, models with more characteristics may not always deliver the highest yield prediction performance. Models
with more and fewer features should be evaluated to discover the best performing model. Several algorithms have been employed in
various research. The findings reveal that while no definitive conclusion can be taken about which model is the best, they do show
that some machine learning models are utilised more frequently than others. The random forest, neural networks, and linear models
are the most commonly utilised models.
REFERENCES
[1] A.T.M.S. Ahamed, N.T. Mahmood, N. Hossain, M.T. Kabir, K. Das, F. Rahman, R.M. Rahman Applying data mining techniques to predict annual yield of
major crops and recommend planting different crops in different districts in Bangladesh 2015 IEEE/ACIS
[2] I. Ahmad, U. Saeed, M. Fahad, A. Ullah, , A. Ahmad, J. Judge Yield forecasting of spring maize using remote sensing and crop modeling (10) (2018)
[3] I. Ali, F. Cawkwell, E. Dwyer, S. GreenModeling managed grassland biomass estimation by using multitemporal remote sensing data—a machine learning
approachIEEE J. 10 (7) (2017),
[4] Cheng, H.; Damerow, L.; Sun, Y.; Blanke, M. Early Yield Prediction Using Image Analysis of Apple Fruit and Tree Canopy Features with Neural Networks. J.
Imaging 2017,
[5] Chen, Y.; Lee, W.S.; Gan, H.; Peres, N.; Fraisse, C.; Zhang, Y.; He, Y. Strawberry Yield Prediction Based on a Deep Neural Network Using High-Resolution
Aerial Orthoimages. Remote Sens. 2019,
[6] Črtomir, R., Urška, C., Stanislav, T. et al. Application of Neural Networks and Image Visualization for Early Forecast of Apple Yield. Erwerbs-Obstbau 54,
69–76 (2012).
[7] Duo Attention with Deep Learning on Tomato Yield Prediction and Factor InterpretationPRICAI 2019: Trends in Artificial Intelligence, 2019,
[8] D. Elavarasan and P. M. D. Vincent, "Crop Yield Prediction Using Deep Reinforcement Learning Model for Sustainable Agrarian Applications," in IEEE
Access, vol. 8, pp. 86886-86901, 2020, S