Shopping.txt file contains fields that indicate whether or not a customer purchased a particular product during a single visit. This file has basic demographics information and their purchase information. 1 means customers did purchase, and 0 means customer did not purchase. I designed this association model (Figure 1) to describe which values of fields typically occur together. This model will help company to manage the inventory and predicting the revenue in the future.
After we input the data, we run the Apriori node in SPSS. Figure 2 (minimum confidence level 80%) gave us overview information about this dataset. The last rule tells us that on 12.341% of the record, alcohol, tinned goods and bakery goods were purchased. Of this group, we see alcohol, tinned goods and bakery goods on 97 shopping trips, and 81.443% also bought ready to make. From this example, we can analyze each rule to get more information.
Then I decrease the minimum confidence level to 75%, 26 rules are produced. By following the rule I listed above, we will be able to find many interesting products are purchased together by customers. After that, we want to test if our model is useful, so we take alcohol as an example. I inserted an alcohol in the rule set text box, and then filtered by customer demographic information. Figure 4 gave us three set of rules whose consequent is buying alcohol. Figure 5 displayed how demographic impact buying alcohol. From figure 5, we learned that rule 3 in figure 4 applies to record 12 of figure 5.
Following these steps, the company can get lot of information from picking up each product as a sample, for example, we can analyze milk and bakery goods. After gathering all these information, company will able to put associative products closer in...