Suppose that you are asked to maintain a very large database containing all transactions of the company. The company relies significantly on mining all these transactions for its future investments. Assume that the minimum support and confidence have already been fixed and they will be the same for all future transactions’ analyses. Your job, in this case, is to mine all the transactions at hand in an efficient and quick manner.

1. What are the advantages of building a data warehouse on top of the database over creating a completely separate data warehouse, in this case? Justify your answer.

2. One efficient way of mining such very large database is to partition your transactions. What is your efficient strategy for partitioning the transaction database?

3. Design a strategy for mining the partitions to obtain global frequent itemsets.

4. Assume that you receive a set of new transactions, every so often. Design a strategy for updating the transaction database or data warehouse.

Answer rating: 100% (QA)

1 Advantages of Building a Data Warehouse on Top of the Database a Data Integration By building a data warehouse on top of the existing database you can integrate data from various sources including t... View the full answer