Forecasting Methods Overview

Forecasting Methods

• Forecasting is a technique for predicting future values based on past observations and relationships between variables.
• Several methods exist, including Time Series forecasting and Associative forecasting.

Time Series Forecasting

• This method relies solely on historical patterns and trends.
• A time series is a sequence of data points collected over time.

Weighted Moving Average

• Similar to a moving average, but assigns more weight to recent values in a time series.
• Different weights can be assigned to different periods in a time series.

Example of Weighted Moving Average Calculation

• Using weights 3, 2, and 1 for the most recent, second most recent, and third most recent periods respectively.
• Example calculations provided with data.

Exponential Smoothing

• A forecasting technique that gives more weight to recent observations, making the method sensitive to trends.
• The formula involves adjusting the previous forecast based on a proportion of the forecasting error.
• Smoothing constant (a) determines the sensitivity of recent changes. A higher value leads to more responsiveness, and a lower value leads to a smoother but potentially slower reaction to fluctuations in values.

Example of Exponential Smoothing Calculation

• Provided example involves a car dealer forecasting March demand based on January (142) and February (153) demand, and an alpha value of 0.2.
• Calculations for the march forecast are demonstrated.

Associative Forecasting

• This method uses known relationships between variables to predict future values.
• Independent variables are factors thought to influence the dependent variable.
• Historical data is needed to assess the relationships.

Linear Regression

• A statistical method to model and analyze the relationship between two or more variables.
• Simple linear regression establishes a linear relationship between one dependent variable and one independent variable.
• Multiple linear regression involves more than one independent variable.
• Used when changes in independent variables predict changes in the dependent variable.

Least Squares Method

• Finding the best-fitting straight line that minimizes the vertical distances between the data points and the line.
• This minimizes the sum of squared errors

Correlation

• Measures the strength of the linear relationship between variables.
• Coefficient of correlation (r) ranges from -1 to +1.
• A value closer to ±1 indicates a stronger relationship; closer to 0 indicates a weaker relationship.

Coefficient of Determination

• (r²) Represents the percentage of change in one variable (e.g., sales) explained or predicted by changes in another variable (e.g., advertising).
• Ranges from 0 to 1. A higher value (closer to 1) indicates stronger explanatory power.

This quiz covers various forecasting methods, with a focus on time series forecasting techniques like weighted moving averages and exponential smoothing. Understand how these methods predict future values based on historical data patterns and relationships. Test your knowledge on calculations and principles behind these forecasting approaches.