Enhancing Agricultural Practices through Machine Learning for Soil

Analysis and Crop Recommendation

Paineni Vaishnavi, Uppara Pavani, Sameena Yousuff, Pitta Pavani and Kalluru Divya Sai

Computer Science and Engineering (Ds), Ravindra College of Engineering for Women, Venkayapally, Kurnool, Andhra

Pradesh, India

Keywords: Crop Recommendation, ML, SVM, DT, Agriculture, Crop Yield Prediction, Soil Analysis, Weather Patterns,

Hybrid Model, Crop Quality Ranking.

Abstract: Based on many factors including Season, Soil type, Rainfall, Temperature, Groundwater Level, Fertilizers,

and Pesticides, the crop recommendation system based on machine learning suggests to the farmer the crop

to be grown. This follows through examination of SVM and DT and a hybrid model of it as a new system on

crop prediction. This is done through crop and crop production data-sets which enable this model to provide

reliable recommendations. It also sorts crops based on quality and its findings help determine the quality of

both high and low-quality leads and helps boost industrial production and economic growth.

1 INTRODUCTION

Agriculture is the backbone of numerous nations,

and India serves as no exception; a large proportion

of individuals in India are engaged in agriculture as

farmers. Traditional agricultural methods based on

knowledge and instinct are not very effective in

determining the category of crops and predicting

production. Machine learning (ML) has recognized

substantial progress in agriculture, enabling data-

driven responses to improve sustainability and yield.

Based on different parameters like soil type,

rainfall, groundwater level, temperature, fertilizers,

pesticides, seasonal conditions, etc., a ML-based

Crop Recommendation System will help farmers

choose the best crop to plant. The system also uses

ML algorithms such as SVM and DT to process large

datasets to give accurate and productive

recommendations. It helps the farmer to get the

maximum yield possible, uses resources efficiently,

and reduces economic loss.

Current crop recommendation systems utilize data

mining to predict weather patterns and agricultural

yield. But these methods, on the other hand, can’t

handle unpredictable temperature and rainfall

patterns that may result in lower accuracy. Moreover,

most of the solutions and implementations are

hardware based and come at a premium cost, and

maintenance proves to be a challenge as well. More

specifically, the proposed system comprises a hybrid

ML model which is a phased model that addresses

issues such as efficiency, reliability in yield

prediction, and crop quality ranking procedures.

This model can help farmers to select the crops

and thus, improve farmers income and economy. In

addition, the system incorporates other modules,

including pesticide prediction and online agricultural

commodity trading, making for a well-rounded

solution to contemporary agricultural difficulties.

2 LITERATURE SURVEY

a) Motwani, Aditya, et al. "Soil Analysis and Crop

Recommendation using Machine Learning."

2023 International Conference for Advancement

in Technology (ICONAT). IEEE, 2023.

https://ieeexplore.ieee.org/document/9725901

India ranks among the world's top three crop

producers. The agricultural economy of India relies on

its impoverished farmers. Soil types vary from area to

region, leaving farmers with limited technical options

when it comes to selecting the most profitable crops

for their soil. Compared to Random Forest's 75%

accuracy, CNN architecture achieves 95.21%.

b) Rao, Madhuri Shripathi, et al. "Crop prediction

using machine learning." Journal of Physics:

568

Vaishnavi, P., Pavani, U., Yousuff, S., Pavani, P. and Sai, K. D.

Enhancing Agricultural Practices through Machine Learning for Soil Analysis and Crop Recommendation.

DOI: 10.5220/0013935300004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 5, pages

568-574

ISBN: 978-989-758-777-1

Conference Series. Vol. 2161. No. 1. IOP

Publishing,2023.

https://www.researchgate.net/publication/35775

9181_Crop_prediction_using_machine_learning

Agriculture is the main income source for most

developing nations. Farming techniques and

agricultural technology are always developing. It is

difficult for farmers to keep up with the demands of

merchants, customers, and the world at large. Soil

erosion and industrial pollution are contributing to

climate change, which farmers must address. (ii)

Phosphorus, potassium, and nitrogen deficiency in the

soil can stunt crop growth. One common error that

farmers do is to produce the same crops every year.

They apply fertilisers carelessly, without knowing the

quantity or quality of the fertiliser they are using. The

goal of the research is to find the most accurate crop

forecast model that can help farmers pick crops

according to weather and soil conditions. Using Gini

and entropy, this study analyses three classifiers:

KNN, DT, and RF. From what we can see, Random

Forest is the most accurate of the three.

c) Priyadharshini, A., et al. "Intelligent crop

recommendation system using machine

learning." 2023 5th international conference on

computing methodologies and communication

(ICCMC). IEEE, 2023.

https://ieeexplore.ieee.org/document/9418375

The agricultural sector plays a crucial role in

India's GDP. In a nation where 58% of the population

works in agriculture, one of the biggest problems is

that farmers often use outdated and unscientific

methods to pick the wrong crops for their land.

Planting crops that aren't well-suited to the soil,

season, and area is a common mistake among farmers.

People end their lives, stop working the land, and go

to cities because of this. To get around this problem,

this study suggests a method that considers all the

variablesto help farmers choose crops. The practice of

precision agriculture, which makes use of modern

agricultural technology to manage crops in a site-

specific manner, is gaining popularity in developing

countries.

d) Pande, Shilpa Mangesh, et al. "Crop

recommender system using machine learning

approach." 2023 5th International Conference on

Computing Methodologies and Communication

(ICCMC). IEEE, 2023.

https://ieeexplore.ieee.org/document/9418351

The majority of rural Indians find gainful

employment in agriculture and related fields. The

country reaps the benefits of its thriving agricultural

sector. Global standards indicate a poor crop output

per acre. The higher suicide rate among marginal

farmers in India might be explained by this. Findings

from this study provide an easy-to-understand and

implement strategy for farmers to predict crop yields.

One possible approach is to use a smartphone app to

link together farmers. Using GPS, the user's location

is ascertained. User enters surface area and soil type.

Algorithms trained by ML select the most profitable

crops and predict farmers' harvests. In order to predict

crop yields, scientists employ SVM, ANN, RF, MLR,

and KNN. At 95% accuracy, Random Forest

outperformed all other methods. In order to maximise

yields, the algorithm also suggests when fertilisers

should be used.

e) Kalimuthu, M., P. Vaishnavi, and M. Kishore.

"Crop prediction using machine learning." 2022

third international conference on smart systems

and inventive technology (ICSSIT). IEEE, 2022.

https://ieeexplore.ieee.org/document/9214190

A certain percentage of domestic production is

provided by agriculture, which is the backbone of

India's economy and ensures food security. But

unnatural climate change is diminishing food

production and forecasting, which is bad news for

farmers' bottom lines since it lowers yields and makes

them less good at predicting crops. This study uses

machine learning, a cutting-edge method for

predicting agricultural yields, to help inexperienced

farmers plant more realistic seeds. The supervised

learning algorithm Naive Bayes recommends it. For

the purpose of assisting their growth, we take readings

of the moisture, humidity, and temperature of

agricultural seeds. An Android app is also in the works

with the software. Users just need to input their current

location and temperature for the program to begin

making predictions.

3 METHODOLOGY

3.1 Proposed System

To make the most of ML for crop selection and yield

prediction, the recommended Crop Recommendation

System examines several factors such as soil type,

rainfall, groundwater levels, temperature, fertilizers,

pesticides, and seasonal situations. Using SVM and

DT algorithms, the system processes large datasets to

provide accurate recommendations, ensuring efficient

resource utilization and increased productivity.

Additionally, a ranking mechanism evaluates crop

Enhancing Agricultural Practices through Machine Learning for Soil Analysis and Crop Recommendation

569

quality, helping farmers distinguish between high and

low-quality yields for better decision-making.

The system also features a crop pest forecasting

and e-commerce module to enhance its utility and

offer a comprehensive approach to modern-day

agricultural challenges. Instead, this model combines

crop datasets with crop production datasets so that it

measures prediction accuracy while minimizing

redundancy. It improves the efficiency of agricultural

and helps the economy reduce the loss of agriculture

and increase its viability and longevity through data-

driven decision-making.

3.2 System Architecture

Crop Recommendation System: Its architecture

(Figure 1) is attached with multi-layers for precise

Crop Selection; Yield prediction. Soil type,

precipitation, temperature, groundwater level,

fertilisers, pesticides, season are just some of the

many variables collected by the Data Collection

Layer. The next step would be processing this data

into the Data Pre-processing Layer which would

include operations such as handling missing values,

normalizing the data, feature extraction, etc.

Next layer is Machine Learning Model Layer where

various models (in our case SVM and DT) will be

used to train and then cross-validate on the historical

crop and production datasets. So, post training, in the

Recommendation & Ranking Layer, it predicts the

most suitable crops according to input parameters and

ranks the crops on their quality. The system provides

some additional smooth features like Pesticide

Prediction and Online Trading Platform which helps

the farmers to get pesticide suggestions and also helps

to make transactions in the market. It's a digital portal

layer that allow farmers to enter data and get

recommended through a mobile or web-based

interactive platform with visual insights on crop

prediction and yield estimation, designing a holistic

digital experience towards advanced agriculture.

Figure 1: Proposed Architecture.

3.3 Modules

3.3.1 Gathering the Datasets

This segment collects crops datasets categorising soil

type, rainfall, groundwater levels, temperature,

fertilisers, pesticides and seasonality from free

resources like Kaggle. After collecting the data, the

proposed model uploads the data for preprocessing

and analysis. With this step, the system receives

reliable, realistic data for correct predictions.

3.3.2 Generate Train & Test Model

Because the ML model is trained on this data, it is

important to clean the data of inconsistencies,

missing values, and to normalise features. The

training and testing datasets are separated with the

80% being training and 20% being testing in the

post-processing. This allows the model to learn from

past data while being able to retain another set that it

will use only for evaluation, improving the fairness of

its predictions.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

570

3.3.3 Run Algorithms

This module will use machine learning models on the

dataset to predict crops. This differs from model to

model of course, but the dataset is typically 70% –

80%training and 30% — 20%testing. The data is up

to October 2023 and is utilized using SVM and DT

algorithms to determine the appropriate crop to be

cultivated against the conditions. These models assist

with the identification of patterns in the data,

enabling better predictions for different farming

conditions. Figure 2 shows the Upload Dataset.

Figure 3 shows the Enter Input Data. Figure 4 shows

the Results.

3.3.4 Obtain the Accuracy

After training and testing the model, this module

measures the accuracy (figure 5) to assess the

performance of the system. The model is evaluated

best on the basis of various metrics like precision,

recall, and F1-score, which indicate how well the

model predicts fitted crops with input variables. In

addition, hyperparameter tuning and more data

preprocessing is done in order to optimize the results,

if accuracy isn't sufficient enough.

3.3.5 Predict Output

The output of the module depends on the parameters

input by the user. By using trained algorithms, we

may estimate the most suitable crop to be planted in

such conditions from these data. The technology can

also rank crops in terms of yield and quality, helping

farmers to make informed decisions. This helps in

providing accurate and dependable suggestions to

maximize the agricultural yield for farmers.

3.4 Algorithms

3.4.1 Decision Tree Classifier

The rule-based DT method divides the dataset based

on feature values for the purpose of decision-making.

Each node in the tree is a choice based on the

parameters input. DT helps in identifying the best

crop in crop recommendation system based on

various parameters like soil type, rainfall, and

temperature. This is an easy to interpret and

computationally efficient algorithm for classification

problems.

3.4.2 Support Vector Machine (SVM)

Support vector machines an SVM is a type of

supervised learning algorithm used for classification.

It finds the best hyperplane to classify data points.

SVM aids in classifying and recommending crops

based on soil conditions, weather patterns, and other

input parameters in this system. Its predictions of crop

selections are accurate, and it is also good with high-

dimensional data.

4 EXPERIMENTAL RESULTS

Accuracy: How well a test can differentiate between

healthy and sick individuals is a good indicator of its

reliability. Compare the number of true positives and

negatives to get the reliability of the test. Following

mathematical:

Accuracy 

  

      

(1)

Precision: The accuracy rate of a classification or

number of positive cases is known as precision.

Accuracy is determined by applying the following

formula:

Precision 



  

(2)

Recall: The recall of a model is a measure of its

capacity to identify all occurrences of a relevant

machine learning class. A model's ability to detect

class instances is shown by percent of correctly

anticipated positive observations relative to total

positives.

𝑅𝑒𝑐𝑎𝑙𝑙 



  

(3)

F1-Score: A high F1 score indicates that a machine

learning model is accurate. Improving model

accuracy by integrating recall and precision. How

often a model gets a dataset prediction right is

measured by the accuracy statistic.

𝐹



𝑆𝑐𝑜𝑟𝑒 

















(4)

𝐹



𝑆𝑐𝑜𝑟𝑒 

    

  

(5)

Enhancing Agricultural Practices through Machine Learning for Soil Analysis and Crop Recommendation

571

Figure 2: Upload Dataset.

Figure 3: Enter Input Data.

Figure 4: Results.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

572

Figure 5: Accuracy Graph.

5 CONCLUSIONS

With the use of machine learning, the proposed Crop

Recommendation System guides farmers towards

optimal crop selection in response to specific agrarian

conditions. By integrating SVM and DT algorithms,

the system improves prediction accuracy and

enhances decision-making in farming. Additionally,

features like crop quality ranking, pesticide prediction,

and an online trading platform provide a

comprehensive solution to modern agricultural

challenges. Not only does this optimize crop choices

but also boosts productivity and economic growth,

ensuring a data-driven, efficient, and accessible

system for farmers.

6 FUTURE SCOPE

Future Scope of Crop Recommendation System:

Integrating with deep learning models such as CNNs

and RNNs for improved accuracy on predictions by

analysing complex agricultural patterns. It assists in

improving soil and weather analysis using IoT-based

sensors for real-time data collection. It would be

beneficial to also include geospatial analysis -- using

things like satellite imagery and GIS -- to determine

land suitability. And the introduction of a mobile app

with multi-language options will provide a more

minimalist approach to recommendations usage by

farmers. Moreover, blockchain technology helps to

be incorporated for secure and transparent

agricultural trading. These innovations will improve

system speed, precision, and user-friendliness, which

will serve farmers and increase crop yield.

REFERENCES

Doshi, Zeel, et al. "AgroConsultant: intelligent crop

recommendation system using machine learning

algorithms." 2018 Fourth International Conference on

Computing Communication Control and Automation

(ICCUBEA). IEEE, 2018.

Kalimuthu, M., P. Vaishnavi, and M. Kishore. "Crop

prediction using machine learning." 2022 third

international conference on smart systems and

inventive technology (ICSSIT). IEEE,2022. M. Kamei,

Trends and Technology In The Era of Post TV-The Rise

of OTT Platforms. Think India Journal, 22(33),

(2019)184-192.

Kulkarni, Nidhi H., et al. "Improving crop productivity

through a crop recommendation system using

ensembling technique." 2018 3rd International

Conference on Computational Systems and Informatio

n Technology for Sustainable Solutions (CSITSS).

IEEE, 2018.

Kumar, Avinash, Sobhangi Sarkar, and Chittaranjan

Pradhan. "Recommendation system for crop

identification and pest control technique in

agriculture." 2019 International Conference on Comm

unication and Signal Processing (ICCSP). IEEE, 2019.

Lacasta, Javier, et al. "Agricultural recommendation

system for crop protection." Computers and Electron-

ics in Agriculture 152 (2018): 82-89.

Motwani, Aditya, et al. "Soil Analysis and Crop

Recommendation using Machine Learning." 2023

International Conference for Advancement in Techno-

logy (ICONAT). IEEE, 2023

Pande, Shilpa Mangesh, et al. "Crop recommender system

using machine learning approach." 2023 5th

International Conference on Computing Methodologie

s and Communication (ICCMC). IEEE, 2023.

Patil, Pavan, Virendra Panpatil, and Shrikant Kokate. "Crop

prediction system using machine learning algorithms."

Int. Res. J. Eng. Technol. (IRJET) 7.02 (2020).

Enhancing Agricultural Practices through Machine Learning for Soil Analysis and Crop Recommendation

573

Priyadharshini, A., et al. "Intelligent crop recommendation

system using machine learning." 2023 5th international

conference on computing methodologies and

communication (ICCMC). IEEE, 2023.

Pudumalar, S., et al. "Crop recommendation system for

precision agriculture." 2016 Eighth International

Conference on Advanced Computing (ICoAC). IEEE,

2017.

Rajak, Rohit Kumar, et al. "Crop recommendation system

to maximize crop yield using machine learning

technique." International Research Journal of Engin-

eering and Technology 4.12 (2017): 950-953.

I.A.Guitart, G. Hervet, S.Gelper, Competitive

advertising strategies for programmatic television. Jo-

urnal of the Academy of Marketing Science, 48,

(2020)753-775. https://doi.org/10.1007/s11747-019-

00691-5

Tengeh, R.K. and Udoakpan, N., 2021. Over-the-top

television services and changes in consumer viewing

patterns in South Africa. Management dynamics in the

knowledge economy, 9(2), pp.257-277.

Van Klompenburg, Thomas, AyalewKassahun, and

CagatayCatal. "Crop yield prediction using machine

learning: A systematic literature review." Computers

and Electronics in Agriculture 177 (2020): 105709.

https://www.kaggle.com/datasets/atharvaingle/crop

recommendation-dataset

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

574