5. Consider the table in the table below. Period Sales (units) January-June 1992 4,075 July-December 1992 8,239 January-June 1993 3,284 July-December 1993 6,194 January-June 1994 2,582 July-December 1994 5,183 January-June 1995 1,968 July-December 1995 4,023 January-June 1996 1,798 July-December 1996 3,522 a. Fit a simple linear regression to this data. What is your forecast for the January-June 1997 period? b. Now divide the data into two sets: the first half of the year, and the second half of the year. Perform a simple regression on the data for sales in the first half-year. What is your forecast for the January-June 1997 period? Do you think this answer is more likely to be accurate than the answer to part (a)? Why? 6. Consider data in problem 5. Now suppose that the publisher has also obtained data on the number of students enrolled in U.S colleges and universities for each of those time periods. Number of Period Sales Students (Units) (millions) January-June 1992 4,075 72 July-December 1992 8,239 71 January-June 1993 3,284 76 July-December 1993 6,194 75 January-June 1994 2,582 74 July-December 1994 5,183 73 January-June 1995 1,968 69 July-December 1995 4,023 68 January-June 1996 1,798 65 July-December 1996 3,522 64 a. Perform a multiple regression on all this data. What is your forecast for the January-June 1997 period? (Assume that the publisher estimates there will be 66 million students enrolled then). b. Perform a multiple regression on the data for sale in the first half-year of each year. What is your forecast for the January-June 1997 period? Do you think this answer is more likely to be accurate than the answer to part (a)? Why?