i need help with questions 3

2. A software quality engineer wants to develop a regression model for predicting the nmber of faults in program () S/He decided that three software complexity metrics are potentially usful as indkpen- dent variables, namely program sine in lines of code in proggr am () program control flow measured in terms of the number o beanches (ra) and the namber of tokens in program ()Data were collected on a set of 24 programs from a distributed application in a data com- munication software system. The ANOVA table for the model +Bu+Baza+Bars+,was found to be as follows Source of Variation d SS MS Model Error Total 3 340989 20 16.3 23 5426.24 (a) Test to determine whether a regression relation exists between and the set of independent variables. Use a level of significance of 05. What do you find? (b) Does the concusion that a regression relation exists imply that the soft ware quality engineer need not screen the independent vari ables and can just use them all? If a regression relation had not existed, what would this imply about screming the independen ariables? 3. Refer to problem (2). The software quality engineer decided to screen (select) the independent variables to determine the best set for predict- ing y. The regression sums of squares for all possible regression models were found to be as follows: Independent variables in moel SSR 2970.61 3654.83 3584.54 5123.80 809 79 3741.30 5409.89 32 (a) For the R2 criterion, indicate which set of independent variables you would recommend as to best set for predicting y (b) An observer states: There are only three variables, so why screen You might as well use all three. Discuss