Linear Regression
Week 9 Summary (DSME5110F)
• Objective: develop a linear equation showing how the response and predictors are related – Key assumptions:
∗ linear relationship (scatter plots)
Copyright By PowCoder代写 加微信 powcoder
∗ error terms (residual plots)
• Idea: Find the linear function that best fits the data
– best: smallest SSE • R syntax: lm()
Simple Linear Regression
• Multiple R-squared
Multiple Linear Regression
• Adjusted R-squared
– Why not multiple R-squared:
∗ adding more predictors can always improve multiple R-squared
∗ Why not more predictors and what are the best predictors (next week)? • Test of Significance
– Overall model: p-value of the F-Statistic – Each predictor: p-value at the end of line
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com