91Ó°ÊÓ

In multiple linear regression, estimating parameters is a crucial step. Parameters define how predictors influence the response variable. In this example, we're interested in the wear of a bearing, influenced by oil viscosity and load.
To start, we need to compute average values for each variable: \( \overline{y} \), \( \overline{x}_1 \), and \( \overline{x}_2 \). With these averages, we calculate the coefficients, or parameters, \( \beta_1 \), \( \beta_2 \,\) and \( \beta_0 \).
This process begins with finding slopes \( \beta_1 \) and \( \beta_2 \) using formulas based on variances and covariances:

• \(\beta_{1} = \frac{\sum (x_{1i} - \overline{x}_{1})(y_{i} - \overline{y})}{\sum (x_{1i} - \overline{x}_{1})^2}\)
• \(\beta_{2} = \frac{\sum (x_{2i} - \overline{x}_{2})(y_{i} - \overline{y})}{\sum (x_{2i} - \overline{x}_{2})^2}\)

Finally, the intercept \( \beta_0 \) is calculated as:

• \(\beta_{0} = \overline{y} - \beta_{1}\overline{x}_{1} - \beta_{2}\overline{x}_{2}\)

These estimates allow us to express the relationship between the response variable and the predictor variables numerically.

Once parameters are estimated, we can predict the response for given predictor values. This uses the estimated regression equation to forecast outcomes. In this task, we're predicting bearing wear based on specific oil viscosity and load.
The prediction formula uses the estimated parameters:

• \( \mu_{Y | x_{1}, x_{2}} = \beta_{0} + \beta_{1}x_{1} + \beta_{2}x_{2} \)

Plugging in the values \( x_1 = 20 \) (oil viscosity) and \( x_2 = 1200 \) (load), you'll substitute these into the equation. This will yield the predicted wear, \( \mu_{Y}\), for those input conditions.
Such predictions are valuable in practical applications, helping to understand how changes in predictors could alter the response.

Response Surface Methodology (RSM) is a collection of statistical techniques useful for modeling and analyzing problems where several variables influence a response of interest. It's widely used in optimizing manufacturing processes and experimental designs.
In the context of multiple linear regression, RSM helps in identifying optimal settings for predictors like oil viscosity and load. These optima point to conditions where the response (wear) is minimized or maximized.
RSM involves:

• Fitting a model that includes linear and potentially quadratic terms.
• Analyzing contours or surfaces of the fitted model to understand variable interactions.

By employing RSM, engineers and scientists can derive conditions that optimize process outputs. This methodology extends basic regression analysis, enhancing its practical applicability.