above picture regressin analysis outcome residual vs fitted value, residual qq-q

Question

above picture regressin analysis outcome residual vs fitted value, residual qq-qplot, standarzied residual vs fitted value, residual vs leverage then solve (d)  

*y-mpg(miles of gallon) x1-wt(weight) x2-sp(speed) x3-vol(cab volume) x4-hp(horse power)

(d) as part of this project, the researcher also wants to find a model that fits the model possibly better than this current model. the researcher is willing to cinsider some potential interactions in the model. describe some possible approaches,being as explocot as reasonably possible

I think you are talking about multi-collinearity issues, for example, whether there is a correlation between horsepower and speed, but I would like to hear from experts.

Fig?re 2 Dingnostie plots for regression of MPG (miles per gallon) of several makes of cars, based on VT (weight), SP (speed); Vot (cab volume) and HP (horse power) 5e El

Explanation / Answer

here we can look the the multicollinearity problem as there may be correlation among the indepenent variables X1,X2,X3,X4.

it is obvious that speed and horse power would be correlated, if it is highly corrleated then we may remove one of the variable and analyze. same idea may be applied any pair of variables.

if indpependet variables ( x1,x2,x3,x4)) are highly correlated to dependent variables Y, then we can take their interaction also. e.g. if X1 and X2 are highly correlated to Y-variable but not much correlated to one-another then we can take their interaction X1*X2 in the model. and same idea may be applied among other pairs or group of variables.

if data would be availble then it can be tried and confirmed by using corrleation matrix of the all the variables(y,x1,x2,x3,x4)

Related Questions

Navigate

• Browse (All)
• Browse A
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.