Electricity-Transformer-Temperature-ETT-Analysis-using-Markov-Transition-Field-MTF-Forecasting-Temperatures-using-ARIMA
STEPS:
1. Import Required Libraries and Packages:
Import necessary libraries and packages for data analysis, visualization, and modeling.
2.Import the Time Series Dataset:
Load the time series dataset(ETTh1 (1).csv) containing 17420 data with timestamp(date) and OilTemperature(OT).
3.Truncate and Plot the DataFrame:
Preprocess the dataset by truncating if needed and visualize the time series data through plots.
4.Discretize the Data:
Discretize the continuous temperature data to convert it into a format suitable for Markov analysis.
5.Create the Adjacency Matrix:
Construct the adjacency matrix representing the relationships between discretized temperature values.
6.Calculate the Markov Matrix:
Calculate the Markov matrix based on transition probabilities between different temperature states.
7.Create the Markov Transition Field:
Utilize the Markov matrix to create a Markov Transition Field, which visually represents the transition probabilities over time.
8.Visualize the Markov Transition Field:
Visualize the Markov Transition Field to observe patterns and properties in the temperature time series data.
9.Downsample the Markov Transition Field:
Optionally downsample the Markov Transition Field for better visualization or computational efficiency.
10.Plot Self-Transition Probabilities on Time Series Data:
Plot self-transition probabilities on the original time series data to gain insights into temporal dependencies.
11.ARIMA Model for Forecasting:
Utilize the ARIMA model for forecasting future temperature values based on historical data.
12.Evaluate Model Performance:
Assess the accuracy of the ARIMA model by calculating metrics such as Root Mean Squared Error (RMSE).