This is an implementation of machine learning for predicting Indonesia Economy Crisis using Random Forest Method
Economic crisis is a severe and sudden upset in any part of the economy. It could be a stock market crash, a spike in inflation or unemployment, or a series of bank failures. They have severe effects even though they don't always lead to a recession. To prevent this, the government must take actions based on several economic variables. The variables provided for this model is :
- Ekspor
- Cadangan Devisa
- IHSG
- Selisih Pinjaman dan Simpanan
- Suku Bunga Simpanan Riil
- Selisih BI Rate Riil dan FED Rate Riil
- Simpanan Bank
- Nilai Tukar Riil
- Nilai Tukar Perdagangan
- M1
- M2/Cadangan Devisa
- M2M
- Krisis (Output/y variable)
After you understands the data and the variables, there is some steps you must take before proceeding to modelling :
- Data Cleaning and Preprocessing
- Data Partition
- SMOTE (Synthetic Minority Oversampling Technique)
In the end, this model generate result :
- 98.45% on Data Testing Accuracy
- 94.44% on Random Forest Accuracy
- etc. (you can look it up in the code file)
The original dataset is from :
https://www.kaggle.com/c/shift-academy-data-science-bootcamp-10/data