Tuscan Lifestyles Customer Strategy
Essay by Siyi Liu • March 19, 2017 • Essay • 607 Words (3 Pages) • 1,059 Views
Assignment #1 Tuscan Lifestyles
Report
Siyi Liu sl888
Part 1
- Logistic regression model
Block 1: Method = Enter
Omnibus Tests of Model Coefficients | ||||
Chi-square | df | Sig. | ||
Step 1 | Step | 6233.253 | 10 | .000 |
Block | 6233.253 | 10 | .000 | |
Model | 6233.253 | 10 | .000 |
Model Summary | |||
Step | -2 Log likelihood | Cox & Snell R Square | Nagelkerke R Square |
1 | 24122.211a | .117 | .258 |
a. Estimation terminated at iteration number 6 because parameter estimates changed by less than .001. |
Classification Tablea | |||||
Observed | Predicted | ||||
Bought "Art History of Florence?" | Percentage Correct | ||||
No | Yes | ||||
Step 1 | Bought "Art History of Florence?" | No | 45126 | 352 | 99.2 |
Yes | 3838 | 684 | 15.1 | ||
Overall Percentage | 91.6 | ||||
a. The cut value is .500 |
Variables in the Equation | |||||||
B | S.E. | Wald | df | Sig. | Exp(B) | ||
Step 1a | last | -.095 | .003 | 1150.401 | 1 | .000 | .910 |
total$ | .001 | .000 | 31.701 | 1 | .000 | 1.001 | |
gender(1) | -.761 | .036 | 452.515 | 1 | .000 | .467 | |
child | -.186 | .017 | 116.097 | 1 | .000 | .830 | |
youth | -.113 | .026 | 18.724 | 1 | .000 | .893 | |
cook | -.270 | .017 | 249.075 | 1 | .000 | .763 | |
do_it | -.539 | .027 | 399.777 | 1 | .000 | .583 | |
refernce | .235 | .027 | 78.087 | 1 | .000 | 1.265 | |
art | 1.156 | .022 | 2723.273 | 1 | .000 | 3.176 | |
geog | .574 | .019 | 950.087 | 1 | .000 | 1.776 | |
Constant | -1.600 | .052 | 943.311 | 1 | .000 | .202 | |
a. Variable(s) entered on step 1: last, total$, gender, child, youth, cook, do_it, refernce, art, geog. |
- Summary and interpretation
- Significance of the model: from the omnibus table, we can see that the Model Chi-Square is statistically significant, given the 0.05 cutoff significance level. We can conclude that the model is statistically significant.
- Model Summary: from the Model Summary table we can see that the Nagelkerke R Square is 0.258, which is pretty good in the case of a logistic regression.
- Percent Correct Predictions: from the Classification table, we can see that the actual and predicted values are summarized in the table with the percentages of correct classifications. Overall, the model correctly predicts or classifies 91.6% of the purchase. The percent of not buy correctly classified is 99.2%, while the percent of buy correctly predicted is 15.1%.
- Variables: from the last table of the logistic regression, we can see that all the variables are significant because all the Sig values are less than 0.05, which is the significance level. The Exp(B) column represents the odds ratio, which measures the effects of the predictors variables. Knowing that positive coefficients will have odds ratio > 1 and negative coefficients will have odds ratio < 1, we can conclude that the total$, refernce, art and geog variables are important. However, the odds ratio of total$ is close to 1, meaning that the coefficient of this variable is near 0. The rest variables have negative coefficients, given that they have odds ratios less than 1.
Part 2 Decile Analysis of Logistic Regression Results
1,2
[pic 1]
3.
Case Summaries | |||
Bought "Art History of Florence?" | |||
Percentile Group of PRE_1 | N | Mean | Sum |
1 | 5000 | .39 | 1935 |
2 | 5000 | .17 | 836 |
3 | 5000 | .10 | 511 |
4 | 5000 | .07 | 368 |
5 | 5000 | .06 | 284 |
6 | 5000 | .04 | 196 |
7 | 5001 | .03 | 139 |
8 | 4999 | .02 | 121 |
9 | 5000 | .02 | 90 |
10 | 5000 | .01 | 42 |
Total | 50000 | .09 | 4522 |
From the table above we can see that the number of customers is 50,000, the number of buyers is 4522, and the response rate of each decile is shown above.
4.
Case Summaries | |||||||||
Mean | |||||||||
Percentile Group of PRE_1 | Total $ spent | Months since last purchase | # purchases, Children's books | # purchases, Youth books | # purchases, Cookbooks | # purchases, Do-it-yourself books | # purchases, Reference books | # purchases, Art books | # purchases, Geography books |
1 | 257.3526 | 7.19 | 1.06 | .51 | 1.07 | .47 | .56 | 1.50 | 1.33 |
2 | 224.8692 | 7.96 | .84 | .39 | .85 | .39 | .40 | .75 | .89 |
3 | 214.2284 | 8.62 | .79 | .37 | .80 | .37 | .38 | .48 | .70 |
4 | 207.6430 | 8.78 | .75 | .36 | .80 | .34 | .31 | .30 | .54 |
5 | 199.1118 | 9.57 | .76 | .33 | .82 | .37 | .27 | .22 | .46 |
6 | 199.1302 | 10.94 | .75 | .36 | .86 | .39 | .26 | .16 | .39 |
7 | 191.3457 | 12.37 | .76 | .35 | .84 | .42 | .23 | .13 | .29 |
8 | 191.5499 | 14.42 | .81 | .36 | .91 | .45 | .21 | .11 | .25 |
9 | 193.6108 | 17.86 | .96 | .41 | 1.12 | .65 | .25 | .13 | .32 |
10 | 204.3416 | 25.87 | 1.07 | .46 | 1.31 | .77 | .25 | .07 | .29 |
Total | 208.3183 | 12.36 | .85 | .39 | .94 | .46 | .31 | .39 | .55 |
...
...