![REFERENCES
Sensors Driven AI-Based Agriculture Recommendation Model for Assessing Land
Suitability by Durai Raj Vincent 1ORCID,N Deepa 1ORCID,Dhivya Elavarasan
1,Kathiravan Srinivasan 1,*ORCID,Sajjad Hussain Chauhdary 2,*ORCID and Celestine
Iwendi 3ORCID
1. Ahmed, a.n.; de hussain, i.d. internet of things (IOT) for smart precision
agriculture and farming in rural areas. IEEE internet things j. 2018, 5, 4890–4899.
[crossref]](https://image.slidesharecdn.com/seminarppt-240629110344-a1d4f438/85/Sensors-Driven-AI-Based-Agriculture-Recommendation-Model-15-320.jpg)
Sensors Driven AI-Based Agriculture Recommendation Model
- 1. Sensors Driven AI-Based Agriculture Recommendation Model for Assessing Land Suitability Department : Information Science And Engineering Name : Pratheeksha R USN : 4NM20IS104 Guide:Mr Vaikunt Pai
- 2. INTRODUCTION Agriculture farming is considered as the base for human living as it provide food,income and employment for most of the countries in the world Agriculture land needs to be assessed before crop cultivation, in order to acquire the properties of the soil, which helps to obtain maximum production One of the causes for the decrease in crop production is the use of the traditional way of cultivation
- 3. In which farmers depend on soil testing labs, which are not able to provide the accurate data The solution for this is replacement with IOT-based sensors ,sensors play a significant role in collecting information about soil parameters With the help of data gathered from different sensors, land suitability analysis could be done, which would help farmers identify the current status of their agriculture land and so that they can improve their crop production Integrating IOT with machine learning models which will provide the farmer recommendation system
- 4. LITERATURE SURVEY This includes the deployment of sensors for real-time data collection The survey highlights the role of wireless sensor networks (WSNS) in modern agriculture. it explores how WSNS enable communication between IOT devices This involves the use of cloud computing technology for storing and processing vast amounts of data generated by IOT devices The survey extensively covers the implementation of AI techniques in agriculture, with a focus on machine learning algorithms It involves the application of Multi Layer Perceptron(MLP), a type of neural network in agricultural for data analysis
- 6. Data collection using sensors : Various sensors, like PH, soil moisture, salinity, and electromagnetic sensors, are placed in farmland,these sensors act like detectives, they collect important information about the soil and its properties Data transmission to cloud via raspberry pi: A raspberry pi, a small computer, collects data from the sensors and sends it to a cloud storage system (AWS) using WI- FI Data processing and storage in the cloud: The data is stored on the cloud, ensuring it's accessible and safe,storing data in the cloud makes it easier to manage and analyze Machine learning model development: An AI model is built using a neural network approach,this model will help in assessing whether the land is suitable for agriculture or not
- 7. Algorithm implementation : An algorithm (MLP)guides the AI model in processing and learning from the vast amount of data Training and assessment of the model:The built model is trained with data from various sensors, and its performance is assessed,this ensures the AI model can accurately classify different types of land based on suitability Classification of land suitability:The model classifies the land into four categories: most suitable, suitable, moderately suitable, and unsuitable,this classification helps farmers make decisions about how to use their land
- 8. IMPLEMENTATION OF THE PROPOSED MODEL 1. Dataset Collection: Dataset includes various parameters like soil texture, granular fragments (indicating sand percentage),soil structure, available water content, porousness, organic matter, PH value, salinity, and carbonates Four decision classes for land suitability assessment: most suitable (class 1) to unsuitable (class 4) 2. Data Preparation: As the data may contain missing and noisy values, the mean of the data is considered Since the data contain different units of measurement (categorical, numerical), normalization is done before applying the proposed model Dataset is then sub-divided into training and independent test sets in the ratio of 75:25 The process of training and testing is repeated for a variable number of iterations until the optimization is met
- 9. 3. Performance measures : To evaluate the performance of the multiclass classification, various metrics are used, such as TP,TN,FP,FN are calculated In the evaluation of multiclass classification, accuracy is a fundamental metric. Accuracy provides an overall assessment of the model's performance ROC-AUC curve provides a performance measure, indicating how well the positive class probabilities are separated from the negative class in various iterations for neural networks (NN) and multi-layer perceptron (MLP) learning models
- 10. AUC-ROC CURVE FOR THE MULTICLASS CLASSIFICATIONS FOR THE NN AND MLP
- 11. EXPERIMENTAL RESULTS The sensor-based AI model assesses agricultural land based on 14 attributes, utilizing a dataset of 1000 instances With 750 instances for training and 250 for testing, the model classifies land into four classes namely most suitable, suitable, moderately suitable, and unsuitable The MLP algorithm is applied to the dataset, and the model's performance is compared with neural networks, exploring various architectural parameters like hidden layers and neurons Results averaged over ten simulations, are presented in tables, showcasing the model's effectiveness under different configurations
- 13. RESULT ANALYSIS The accuracy of the results is determined by how well the NN and MLP models can classify data into different categories The performance of neural network (NN) training changes with different Nh values (e.g., Nh =30, Nh = 50, Nh = 80). as Nh increases, the NN performs better, accuracy increase with an increase in Nh MLP with three hidden layers are found to be much better than that of NN. similar to the NN model, this MLP model shows improved performance results with the increasing number of Nh The accuracy and other performance measures are found to improve accordingly with an increase in the Nh . on observing the performance measures of MLP with four hidden layers, the model individually is found to provide better results with improved performance with an increase in Nh Nh = 30, MLP with three hidden layers is found to produce better results than the MLP with four hidden layers, The performance of MLP with four hidden layers is high compared to the performance of the neural network and the MLP with three hidden layer approaches
- 14. CONCLUSIONS Agriculture, as a country's backbone, requires sustained growth, and the model presented in this work aims for an impressive 99% accuracy The efficient handling of data from various sensors, managed through an MLP with four hidden layers, ensures enhanced agricultural practices The incorporation of a precise advisory system contributes to better results in farming The proposed approach's high accuracy and precision, along with multiclass classification, offer a sophisticated tool for farmers' guidance and improved real-time decision-making for better crop yield productivity
- 15. REFERENCES Sensors Driven AI-Based Agriculture Recommendation Model for Assessing Land Suitability by Durai Raj Vincent 1ORCID,N Deepa 1ORCID,Dhivya Elavarasan 1,Kathiravan Srinivasan 1,*ORCID,Sajjad Hussain Chauhdary 2,*ORCID and Celestine Iwendi 3ORCID 1. Ahmed, a.n.; de hussain, i.d. internet of things (IOT) for smart precision agriculture and farming in rural areas. IEEE internet things j. 2018, 5, 4890–4899. [crossref]