In: Statistics and Probability
The following data give the average price received by fishermen for several species of fish in 2000 and 2010. The price is in cents per pound.
Fish | Year 2000 Price (x) | Year 2010 Price (y) |
---|---|---|
COD | 13.1 | 56.0 |
FLOUNDER | 15.3 | 166.7 |
HADDOCK | 25.8 | 105.5 |
MENHADEN | 1.8 | 41.3 |
PERCH | 4.9 | 104.2 |
CHINOOK | 55.4 | 236.8 |
COHO | 39.3 | 135.6 |
ALBACORE | 26.7 | 84.6 |
SOFT SHELLED CLAMS | 47.5 | 222.6 |
LOBSTERS AMERICAN | 94.7 | 374.7 |
SEA SCALLOPS | 135.6 | 432.6 |
SHRIMP | 47.6 | 225.4 |
a.
Here, the explanatory variable is 2000 prices and the response variable is 2010 prices.
Let the regression line be of the form,
y=β0+β1x+e,
where β0 is the shift from the origin orintercept and β1 is the
slope of the equation and e is the error
and e is the error
Let the ith observation be,
yi=β0+β1xi+ei, here i=1(1)12
Now, we want to minimize the error sum of square i=112ei2.
So we solve the normal equations for α and β and get,
β1=covx,ySx2 and β0=y-β1x
X=1ni=1nxi, Y=1ni=1nyi
Sxx=1n-1i(xi-x)2=sample variance of x
Syy=1n-1i(yi-y)2= sample variance of y
covx,y=1nixiyi-xy= population covariance b/w x and y.
Sxy=1n-1ixiyi-xy=sample covariance b/w x and y
All the calculations are done using excel functions.
Now, the calculations are shown below,
x_bar |
42.301 |
y_bar |
182.17 |
Sxx |
1529.3 |
Syy |
15084 |
Sxy |
4569 |
Now, we estimate the parameters of the equation by least square method and have,
β0 |
55.76 |
β1 |
2.98 |
Therefore, the required regression equation is,
y year 2000=55.76+2.98x (year 2010)
b.
The correlation coefficient, r=cov(x,y)SxxSyy
Therefore, the correlation coefficient 45691529.3×15084 =0.9513
c.
Now, we have to find the expected price of a fish which was priced 41.3 cents per pound in 2000.
We replace x=41.3 in the regression equation and have,
The predicted price of the fish in 2010 (y) = 55.76+(2.98*41.3) = 178.834 cent per pound.