Integrated Crop Recommendation System: Ensuring Confidentiality of Data
Siddharth Sharma | GUIDE: Parthasarathy G | SCOPE
Introduction Results
Agriculture, the fundamental source of worldwide nourishment, confronts a Evaluation Metrics
changing terrain characterized by dynamic environmental factors and the Models Accuracy Precision Recall F1-Score
essential need for sustainable practices. The precision of crop selection in
Decision Tree 0.8926 0.8964 0.8936 0.8852
response to varying conditions has become paramount for ensuring food security Gaussian Naïve 0.9443 0.9438 0.9443 0.9425
and optimizing resource utilization. Traditional crop recommendation systems, Bayes
although valuable, often encounter challenges in providing accurate and reliable SVM 0.9443 0.9461 0.9443 0.9408
suggestions due to the intricate nature of modern agricultural decision-making. Logistic 0.9131 0.8975 0.9131 0.9014
Regression
Random Forest 0.9215 0.9221 0.9216 0.9217
Motivation KNN 0.9216 0.9221 0.9216 0.9218
Enhancing the accuracy of the existing algorithms to provide more precise Gradient 0.9215 0.9221 0.9215 0.9217
recommendations for crop selection. Providing confidentiality of data with the Boosting
help of encryption and decryption. Algorithms for selection are chosen at random User Interface (GUI)
from a list of algorithms.
SCOPE of the Project
Data Collection and Preprocessing: Gather data on crops, soil properties (pH, GUI Output
nutrients), weather patterns (temperature, rainfall), and historical yields. Clean
and format the data for machine learning.
Model Development: Choose and train a machine learning model (e.g., Random
Forest, Logistic Regression) to analyze the data and predict suitable crops based
on user inputs.
System Design and Development: Design a user interface (web app, mobile app)
for farmers to interact with the system. This interface should allow users to input
data about their land and receive crop recommendations.
Target audience: Tailor the system's complexity and user interface to the farmers'
technical proficiency (basic vs. smartphone-savvy).
API Inputs
Scalability: Design the system to handle a large number of users and data inputs if
intended for a wide geographical region. API Outputs
Data privacy: Ensure user data (land details, harvest information) is stored
securely if the system collects such information.
Expand data inputs: Include factors like water availability, historical pest
problems, and market prices for different crops.
Real-time data integration: Integrate weather sensors or market data feeds to
provide up-to-date recommendations.
Methodology Conclusion
The dataset is first collected, preprocessed, and correlation is found. •The crop recommendation system will analyze soil and climate data to provide
Based on that it is also checked if the dataset is biased or unbiased. accurate recommendations for suitable crops.
The dataset is then split into training and testing data and trained with different •The system considered various factors such as temperature, rainfall, soil pH, and
ML models like Decision Tree, Support Vector Machines, Random Forest, Gradient nutrient levels to make informed recommendations.
Boost, Naïve Bayes, Logistic Regression. •The data was successfully sent and received while maintaining confidentiality of
Selecting all the models and check the output of each of them. data.
Then select the output as result which is present in majority of the algorithm. • We have received outputs in the form of GUI as well as JSON format for further
The model is saved as a .pkl file for use in the encryption decryption program. usage.
The user initiates a connection to the server through socket programming.
The server generates a secret key to share it with the user via socket connection. References
Once the key is received the user can access the GUI/ API [1]Bandara, P., Weerasooriya, T., Ruchirawya, T., Nanayakkara, W., Dimantha, M.,
The user enters the details in GUI or in JSON format. & Pabasara, M. (2020). Crop recommendation system. International Journal of
Computer Applications, 975, 8887.
The data is then encrypted using any algorithm out of TDES, CAST, Blowfish, AES.
[2]Pudumalar, S., Ramanujam, E., Rajashree, R. H., Kavya, C., Kiruthika, T., & Nisha,
The encrypted data is sent to the server end where it is decrypted. J. (2017, January). Crop recommendation system for precision agriculture. In 2016
The decrypted data is sent to all the saved ML models. Eighth International Conference on Advanced Computing (ICoAC) (pp. 32-36). IEEE
The output of each model is taken in to consideration and the majority of the [3]Rajak, R. K., Pawar, A., Pendke, M., Shinde, P., Rathod, S., & Devare, A. (2017).
output is taken as the final output. Crop recommendation system to maximize crop yield using machine learning
The result is then sent back to the sender side as a response. technique. International Research Journal of Engineering and Technology, 4(12),
The response is displayed in a dialog box in case of GUI and a JSON output in case 950-953.
of API [4]Gosai, D., Raval, C., Nayak, R., Jayswal, H., & Patel, A. (2021). Crop
recommendation system using machine learning. International Journal of
Scientific Research in Computer Science, Engineering and Information Technology,
7(3), 558-569.
[5]Lacasta, J., Lopez-Pellicer, F. J., Espejo-García, B., Nogueras-Iso, J., & Zarazaga-
Soria, F. J. (2018). Agricultural recommendation system for crop protection.
Computers and Electronics in Agriculture, 152, 82-89.
[6]Reddy, D. A., Dadore, B., & Watekar, A. (2019). Crop recommendation system to
maximize crop yield in ramtek region using machine learning. International
Journal of Scientific Research in Science and Technology, 6(1), 485-489.
[7]Kumar, A., Sarkar, S., & Pradhan, C. (2019, April). Recommendation system for
crop identification and pest control technique in agriculture. In 2019 International
Conference on Communication and Signal Processing (ICCSP) (pp. 0185-0189).
IEEE.
