Temperature Forecast - ACSE-5

Usage

Download

To download the data files from Met Office, make sure you have installed Curl (or from GitHub) on your device. Then, compile the DataUtil.cpp source file linked with curl.

g++ DataUtil.cpp -o Download.out -lcurl && ./Download.out

Prediction using FFT

To compile the main file, using

g++ Main.cpp -o Main.out

The program takes two arguments as input, the station name and the periods you wanna predict.

For example, if you want to predict the average temperatures of next 12 months of Heathrow Station, the command should be

./Main.out heathrow 12

Result

Method 1 - Fast Fourier Transformation - BETTER

• Heathrow Station - Next 5 Years - RMSE: 0.0544736

• Shawbury Station - Next 10 Years - RMSE: 0.118409

Method 2 - Exponential Smoothing - BAD

• Basic Exponential Smoothing RMSE: 3.15936

• Double Exponential Smoothing RMSE: 3.08166

• Triple Exponential Smoothing RMSE: 3.18623

Data Source

• Met Office Climate and climate change - Historic station data