Cluster analysis of SPSS is instance
(1) the survey parameter
Drinks drink sugar price number heat net content color palate quality month several factors influencing the consumers to buy
1.00 56.00 450.00 1.00 12.00 70.00 12.00 (1) net content of 1
2.00 78.00 500.00 5.00 15.00 80.00 12.00 1 (2) sugar, shelf life
3.00 45.00 250.00 2.00 19.00 2, 60.00 6.00 (3) heat
Price of 4.00 50.00 12.00 106.00 550.00 10.0025.00 4 (4)
70.00 12.00 5.00 245.00 500.00 11.0019.00 3 (5) color
6.00 65.00 250.00 1.00 8.00 90.00. 1 50 6 flavors
7.00 55.00 300.00 2.00 4.00 40.00 1.00 1
8.00 60.00 250.00 2.00 2.50 90.00 6.00 1
2 50.00 12.00 9.00 182.00 470.00 12.005.00
10.00292.00 15.008.00 3 90.00 12.00 500.00
11.00140.00 10.0015.00 1 95.00 24.00 550.00
12.0075.00 250.00 1.00 2.00 40.00 12.00 3
13.0044.00 255.00 2.00 4.00 2, 50.00, 6.00
00 14.0021.00 50.00. 6.00 2, 60.00 8.00
15.0015.00 200.00. 00 1 2.00 80.00. 30
16.00189.00 450.00 5.00 4.50 75.00 12.00 4
(2) the results

One-way ANOVA
Clustering error significance F
The mean square df mean square df
51529.0001 3562.429 14.465, 14 002
Net content of 292951.5631 3647.098, 80.325, 14 000
Sodium 1, 10.813 217.563 20.121, 14 001
Price 196.000 1 41.103 4.769, 14 046
Color 2.250 1 1.125 2.000, 14 179
Taste is 306.250 1 349.107 14. 877. 365
Shelf life 290.702 117.524 1416.589 months. 001
Only when should be used to describe the purpose F test, because has chosen clustering is used to maximize different clustering in the case of difference, therefore be corrected by the observation of the significant level has not been, so I can't interpret these levels as "the clustering method are equivalent" hypothesis test,




(3) the results analysis (this data is not the actual collection, the results may also be inconsistent with the actual)
When the relative distance between the types of 5, 16 kinds of drinks can be divided into 9 classes;
When the relative distance between the types of 10, 16 kinds of drinks can be divided into five types;
And types of the relative distance between 25, can only be classified as a category;
So, the influence factors of different beverage sales for 25 units (unknown),
The significant data shows that consumers buy drinks, drinks consider net content is given priority to, shelf life for the time, the rest to see prices and calories, and then choose their favorite flavors and colors,
That is, the content of drinks popular point, content should be in line with the price; Shelf life should is that consumers will see the link, expired or is about to expire (one month) will lose consumer; Experience after finally choose to drink drink, ha ha!