91Ó°ÊÓ

Linear regression involves finding the best fit line through a set of data points using the least squares method. This is essentially about discovering the slope and intercept of a straight line, which best capture the trend in the data. The model is typically represented as:

• \( y = ax + b \)

In this formula, \( a \) is the slope of the line, while \( b \) is the y-intercept. The method of least squares aims to calculate \( a \) and \( b \) so that the sum of the squares of the vertical differences, or residuals, between the actual and estimated values is minimized.The formulas used in linear regression to calculate \( a \) and \( b \) are as follows:

• \[ a = \frac{n\sum{x_iy_i} - \sum{x_i}\sum{y_i}}{n\sum{x_i^2} - (\sum{x_i})^2} \]
• \[ b = \frac{\sum{y_i} - a\sum{x_i}}{n} \]

By applying these formulas, we can determine a linear relationship that predicts dependent variable values from an independent variable, capturing a trend over the data span.

Quadratic fit takes data analysis a step further by allowing the model to curve, rather than just follow a straight line. This is especially useful when data shows a parabolic trend. The quadratic model we fit is:

• \( y = ax^2 \)

In this case, we are looking for the coefficient \( a \) that defines how the curve fits the data. Using least squares, we find the best value of \( a \) by minimizing the sum of the squared differences between observed values and those predicted by our model.The calculation of \( a \) is given by the formula:

• \[ a = \frac{\sum{x_i^2 y_i}}{\sum{x_i^4}} \]

Through this approach, the quadratic fit can capture and adapt to nonlinear patterns, providing a more flexible model when comparing against real-world data that's not perfectly linear.

A cubic fit is used when data follows a more complex pattern which cannot be effectively modeled by linear or quadratic equations. Here, the model follows a cubic equation:

• \( y = ax^3 \)

The method of least squares is again applied, helping us to find the coefficient \( a \) that minimizes the residuals. The main advantage of cubic fits is their ability to model more inflections and changes in direction due to their additional degree.The formula used to compute \( a \) in a cubic fit is:

• \[ a = \frac{\sum{x_i^3 y_i}}{\sum{x_i^6}} \]

This type of fit is instrumental when looking at datasets with varying trajectories, making it a powerful tool in capturing more sophisticated data trends without being overly complex.

Polynomial models expand our fitting capabilities further, by combining terms of different powers into a single model. This approach is ideal when data displays intricate patterns that linear, quadratic, or even cubic fits cannot capture.For instance, a polynomial model may look like:

• \( y = ax^3 + bx^2 + c \)

Here, the least squares method finds coefficients \( a \), \( b \), and \( c \) that minimize the discrepancy between the data and the model. This involves solving a system of equations known as the normal equations:

• \( \sum{y_i} = c n + b \sum{x_i^2} + a \sum{x_i^3} \)
• \( \sum{x_i^2 y_i} = c \sum{x_i^2} + b \sum{x_i^4} + a \sum{x_i^5} \)
• \( \sum{x_i^3 y_i} = c \sum{x_i^3} + b \sum{x_i^5} + a \sum{x_i^6} \)

By fitting these coefficients, polynomial models capture the nuanced features of complicated datasets, allowing for intricate modeling of real-world phenomena.