IT 9626 - Modelling

Objective: Apply modelling to areas (financial forecasting, climate change)

What is Modelling?

Modelling is the process of creating a simplified representation of a real-world system or phenomenon. This representation, or model, allows us to understand, analyze, and predict the behavior of the real system. Models can be mathematical, physical, or computational. In IT, we often use computational models to simulate real-world scenarios.

Types of Models

There are various types of models, each suited for different purposes:

• Conceptual Models: High-level representations using diagrams (e.g., UML diagrams) to illustrate system components and their relationships.
• Mathematical Models: Using equations and mathematical principles to describe the system's behavior.
• Simulation Models: Computer programs that simulate the system's operation over time.
• Statistical Models: Using statistical techniques to analyze data and make predictions.

Modelling in Financial Forecasting

Financial forecasting involves predicting future financial performance. Models play a crucial role in this process.

Time Series Analysis

Time series analysis uses historical data points collected over time to forecast future values. Common time series models include:

• ARIMA (Autoregressive Integrated Moving Average): A statistical model that uses past values to predict future values.
• Exponential Smoothing: A technique that assigns exponentially decreasing weights to past observations.

Regression Models

Regression models establish relationships between dependent and independent variables. For financial forecasting, this could involve predicting stock prices based on economic indicators.

Model	Description	Example Application
ARIMA	Uses past values of a time series to predict future values.	Predicting future sales based on historical sales data.
Exponential Smoothing	Assigns exponentially decreasing weights to past observations.	Forecasting demand for a product.
Linear Regression	Models the relationship between a dependent variable and one or more independent variables.	Predicting stock prices based on economic indicators like interest rates and inflation.

Modelling in Climate Change

Climate change models are complex computational models used to simulate the Earth's climate system and predict future climate scenarios.

Climate Models (GCMs)

Global Climate Models (GCMs) are sophisticated computer programs that divide the Earth's atmosphere and oceans into a three-dimensional grid. These models use physical laws (e.g., thermodynamics, fluid dynamics) to simulate the interactions between different components of the climate system.

Key Variables in Climate Models

Climate models simulate various key variables, including:

• Temperature: Average temperature at different altitudes.
• Precipitation: Rainfall, snowfall, and other forms of precipitation.
• Wind: Wind speed and direction.
• Sea Level: Changes in sea level.
• Greenhouse Gas Concentrations: Concentrations of gases like carbon dioxide and methane.

Model Validation

Climate models are constantly being validated against historical data to ensure their accuracy. This involves comparing model outputs with observed climate changes.

Scenario Analysis

Climate models are used to explore different future climate scenarios based on various assumptions about greenhouse gas emissions and other factors.

Benefits of Modelling

Using models offers several benefits:

• Understanding Complex Systems: Models simplify complex systems, making them easier to understand.
• Prediction: Models can be used to predict future behavior.
• Decision Making: Models provide insights that can inform decision-making.
• Risk Assessment: Models can help assess potential risks.

Limitations of Modelling

It's important to recognize that models are simplifications of reality and have limitations:

• Simplifications: Models inevitably simplify real-world systems, potentially omitting important details.
• Assumptions: Models rely on assumptions, which may not always be accurate.
• Data Requirements: Building and validating models often require large amounts of data.
• Computational Complexity: Some models can be computationally expensive to run.