HW 5 – Shopping Cart Analysis
final project代写 The primary mission is to build the Apriori association rule model of retail commodities and analyze the correlation between them. The…
The primary mission is to build the Apriori association rule model of retail commodities and analyze the correlation between them. The information of one attribute can be used to infer another attribute, and it will be validated when the confidence reaches a certain threshold.
-
Define the index final project代写
○ Support – the probability that two items ( ) appear in the total number of sales (N), that is, and the probability that they are purchased at the same
time.
○ Confidence – a conditional probability that represents the probability of purchasing B after purchasing A.
○ Lift – it should refer to the effect of purchasing first on the probability of purchasing.
■ Comparison of two possibilities, one is the probability of buying the right item given that the item on the left is known to be purchased, and the other is the probability of buying the item on the right under any circumstances.
■ Only when the value of the lift (A->B) is greater than 1 can the association rule be truly effective, indicating that the more item A is sold, the more item B will be sold. A lift equal to 1 means that there is no association between products A and B. Finally, if the lift is less than 1, it means that buying A will reduce the sales of B.
-
Apriori algorithm to generate frequent itemsets final project代写
○ In this stage, all itemsets that satisfy the minimum support are found, and these itemsets found are called frequent itemsets.
○ If an item set is frequent, all its subsets are also frequent.
○ If a subset is infrequent, then all itemsets that contain it are infrequent.
- Obtain association rules through credibility filtering in frequent sets. Then build a recommendation system based on association rules, record the top 10 product combinations.
The final product should find the association between different commodities in store, and formulate sales strategies according to the association rules between commodities.
First, mining out all support and confidence greater than a certain threshold from the shopping cart data set. The result will be present when X and Y are related, and users who bought X are likely to buy Y. E.g. :
{yogurt, vegetable}→ {milk} support:0.0200000, confidence: 0.500000, lift: 2.000000
{milk, yogurt} → {vegetable} support:0.020000, confidence: 0.300000, lift: 2.000000
According to the analysis, customers will most likely purchase milk while purchasing yogurt. And vegetables, the maximum confidence level is 50%. Then, form a recommendation system after the association mining to have a better demonstration of which product has a high degree of association for the store keeper.
