This repository was created to accommodate the Capstone (Taniland) project model in the Wake 2023 Machine Learning Path program, which is a land management application aimed at the agricultural industry so that it can assist in better land management through the features we have made in this application.
To make a plant recommendation feature we create a model through temperature, humidity and image data according to land conditions, in making this feature we use 3 algorithms to build a planting recommendation model:
- by IoT
- Clustering : Kmeans
- Classification : Decission Tree
- by Image
- Deep Learning : CNN
Steps to run this project:
- Crop Recomendations
- Clustering :
- Upload crop_recomendation.csv into notebook
- Run runtimes
- The results of each cluster will become a new dataset
- Export the new dataset to a .csv file for classification using a decision tree
- The model will be saved in .pickle format
- Klasifikasi :
- Train each new cluster dataset into a decision tree model
- The model will be saved in .pickle format
- Soil type prediction
- Enter data into Google Drive
- Customize the data train path
- Run runtimes
- The model will be saved in .h5 and .json format