Skip to content

Latest commit

 

History

History
28 lines (21 loc) · 780 Bytes

README.md

File metadata and controls

28 lines (21 loc) · 780 Bytes
Raw

Machine-Learning-ICA

To create a Electricity Demand forecast for France on an hourly basis up to 48 hours ahead.

Dataset given had 5 years of historical data.

Used Regression and Time-series analysis.

**Time Series Algorithm used: **

  1. Seasonal ARIMA.

Regression Algorithms used :

  1. Linear Regression.
  2. K Nearest Neighbors (KNN) Regressor.
  3. Random Forest Regressor.
  4. XGBoost Regressor.
  5. AdaBoost Regressor.
  6. Decision Tree Regressor.

Accuracy

ARIMA (1,0,1) (1,0,1,12) - 98.92% XGBoost Regressor - 97.53% Random Forest Regressor - 96.62% K Nearest Neighbors Regressor - 95.06% Decision Tree Regressor - 92.90% AdaBoost Regressor - 82.90% Linear Regression - 72.20%