![Grid Search Optimization
Hyperparameters search space The best hyperparameters with GS
algorithm
'max_depth': [80, 90, 100], 'max_depth': 80,
'max_features': [2, 3], 'max_features': 3,
'min_samples_leaf': [3, 4, 5], 'min_samples_leaf': 3,
'min_samples_split': [8, 10, 12], 'min_samples_split': 8,
'n_estimators': [100, 200, 300] 'n_estimators': 200](https://image.slidesharecdn.com/masterpresentation-241027081745-a5c71714/85/Design-Smart-Agriculture-System-using-IoT-Technology-20-320.jpg)
Design Smart Agriculture System using IoT Technology
- 1. University of Technology control and systems Engineering Department Design smart Agriculture system using Iot Technology Dunia Gzar Supervisors Dr. Ali Majeed & Dr. Maythem Kemal
- 3. Contents: Problem statement Aims& Objective Systems proposal Methodology Results publication
- 4. Problem Statement 1.Increasing Population 2.Traditional agriculture 3. Immigration 4.Weather challenging
- 5. Aims 1. Real time data 2. Remotely monitoring and controlling 3. Acceleration decision making 4. Increasing productivity while reducing manpower (cost , resources, safety)
- 6. Objectives 1. Reviewing the related works. 2. Studying the types of IoTsensors. 3. Designing an IoT hardware and software components. 4. Improving the system design via the application of proper machine learning techniques 5. Optimization technique is involved such as Grid Search
- 7. Materials of the proposal system consist of 1. ESP32 2. Humidity and temperature sensor 3. Soil moisture sensor 4. Gas sensor 5. Rain sensor 6.Light intensity 7. NPK Sensor with Arduino
- 8. Flowchart of General Operation of proposal system 8
- 9. Schematic Proposal Design 9 NPK sensor gas sensor Rain sensor DHT22 Light intensity ESP32
- 10. ESP32 10
- 11. NPK Sensors 11 Soil Moisture Sensor, 5 Pin Comprehensive Sensor Corrosion Resistant Probe, for Soil pH + NPK + Temperature and Humidity + Conductivity Detection, for Various Soils Multi-function High Accuracy High Quality Good Airtightness
- 12. NPK Sensor mapping with Arduino 12 https://i0.wp.com/www.electroniclinic.com/wp-content/uploads/2021/03/Soil-NPK-Sensor- interfacing-with-Arduino-circuit-diagram-including-bluetooth-module-and-oled-display.png?ssl=1
- 13. IoT Sensors 13
- 14. Amazon Web Service 14 AWS has more services, and more features within those services, than any other cloud provider, including compute, storage, databases, networking, data lakes and analytics, machine learning and artificial intelligence, IoT, security, and much more
- 15. Collecting Data Using IoT Sensors 15
- 16. Compare R2, MAE, MSE 16 Data Linear Regression Algorithm SVR Algorithm Random Forest Algorithm MLP Algorithm Methodology
- 18. Results of Data Processing With ML algorithms Algorithms MSE MAE R2 Score LR 0.0031 0.039 0.943 RFR 0.0002 0.005 0.995 MLP 0.0032 0.037 0.942 SVR 0.0034 0.045 0.939
- 19. Plotted Results
- 20. Grid Search Optimization Hyperparameters search space The best hyperparameters with GS algorithm 'max_depth': [80, 90, 100], 'max_depth': 80, 'max_features': [2, 3], 'max_features': 3, 'min_samples_leaf': [3, 4, 5], 'min_samples_leaf': 3, 'min_samples_split': [8, 10, 12], 'min_samples_split': 8, 'n_estimators': [100, 200, 300] 'n_estimators': 200
- 21. Results Model MSE MAE R2 Time processin g RF 0.000 3 0.007 0.954 1.253 RF/ Hyper tuned 0.0002 0.004 0.998 1.61 Improvement 33 % 44% 5 %
- 23. Next Step#1
- 24. Next step #2 Caching layer Sensor node Decreased network costs Improved responsiveness Increased performance on the same hardware Availability of content during network interruptions Sensor node