metadata
title: Transformer Model for Energy Consumption Prediction
description: >-
This model predicts energy consumption based on meteorological data and
historical usage.
license: gpl
Transformer-Architecture for Energy Consumption Prediction
Description
This model applies Transformer architecture to predict energy consumption over a 48-hour period using historical energy usage and weather data from 2021 to 2023.
Model Details
Model Type: Transformer
Data Period: 2021-2023
Variables Used:
- Transformer with Energy consumption data and weather data
- Transformer with Energy consumption data and two additional variables: 'Lastgang_Moving_Average' and 'Lastgang_First_Difference'
Features
The model uses a sequence length of 192 (48 hours) to create input sequences for training and testing. Positional encodings are added to the sequences to provide temporal information to the Transformer model.
Installation and Execution
To run this model, you need Python along with the following libraries:
pandas
numpy
matplotlib
scikit-learn
torch
gputil
psutil
torchsummary
Steps to Execute the Model:
Install Required Packages
Load your Data
Preprocess the data according to the specifications
Run the Script