Hello, I am trying to teach myself about UML case diagrams. I've read some documentation and found this example online and have attempted my own solution. I was wondering if anyone could validate it? My solution can be found at: http://img176.imageshack.us/my.php?image=usecasedd2.jpg
The following is what I used to create the case:
After several meetings, it is determined that you are to develop a data mining system design for the company. Vanessa excitedly led her team to meet with you multiple times to discuss their expectations. Here is a brief description:
“The system should be strictly a back-office application. To begin with, whenever a customer registered through the web site or regular mail, the marketing department will enter his demographic information into the customer database. As soon as that happens, we want the data to flow into our database automatically.
Each transaction would be carried out by the customers and would be captured by the real-time transactional system monitored by IT. We want that data flow to come into our database real time as well.
Data miners (that is us) constantly monitor all activities through this system. We conduct many different analyses. First, we cleanse the data. There are potential “noises” in the data. Dirty data caused by human error when they are entered, bad data like canceled purchase, etc. We start by setting up several flags in the system such as illegal data type, bad dates, marker for cancellation, etc. The system should then scan the data, filtering out all the bad data that raise flags.
Once the data is cleansed, we start several analyses. The CRM analysts will conduct clustering analysis based on demographic information and spending patterns. Specifically, they will retrieve customer’s name, gender, age, income, zip code, total amount of spending in the last month, spending breakdowns, and try to form “clusters” of different customers for future marketing campaign use. A report is printed out monthly to describe the landscape of our customer base and sent to the marketing department.
The Fraud analysts, on the other hand, focus on fraud detection. They first retrieve customerÂ’s name, gender, age, income, zip code. Then they retrieve each transaction over the past year for the customer. Each transaction should contain total amount charged, status (paid, being processed, canceled), location, store, date and time. Based on all this information, the analysts would then use statistical outlier analysis to find abnormal transactions. If no abnormal transactions are found, a sophisticated neural network program would be ran to further identify hard-to-crack fraud events.
When fraud is detected, the analysts will press a button in the system. An email with the customer and credit card information would be generated automatically by the system and sent to the marketing department for immediate handling.
All analysts need to create an account and log in before they can use the system.
The following is what I used to create the case:
After several meetings, it is determined that you are to develop a data mining system design for the company. Vanessa excitedly led her team to meet with you multiple times to discuss their expectations. Here is a brief description:
“The system should be strictly a back-office application. To begin with, whenever a customer registered through the web site or regular mail, the marketing department will enter his demographic information into the customer database. As soon as that happens, we want the data to flow into our database automatically.
Each transaction would be carried out by the customers and would be captured by the real-time transactional system monitored by IT. We want that data flow to come into our database real time as well.
Data miners (that is us) constantly monitor all activities through this system. We conduct many different analyses. First, we cleanse the data. There are potential “noises” in the data. Dirty data caused by human error when they are entered, bad data like canceled purchase, etc. We start by setting up several flags in the system such as illegal data type, bad dates, marker for cancellation, etc. The system should then scan the data, filtering out all the bad data that raise flags.
Once the data is cleansed, we start several analyses. The CRM analysts will conduct clustering analysis based on demographic information and spending patterns. Specifically, they will retrieve customer’s name, gender, age, income, zip code, total amount of spending in the last month, spending breakdowns, and try to form “clusters” of different customers for future marketing campaign use. A report is printed out monthly to describe the landscape of our customer base and sent to the marketing department.
The Fraud analysts, on the other hand, focus on fraud detection. They first retrieve customerÂ’s name, gender, age, income, zip code. Then they retrieve each transaction over the past year for the customer. Each transaction should contain total amount charged, status (paid, being processed, canceled), location, store, date and time. Based on all this information, the analysts would then use statistical outlier analysis to find abnormal transactions. If no abnormal transactions are found, a sophisticated neural network program would be ran to further identify hard-to-crack fraud events.
When fraud is detected, the analysts will press a button in the system. An email with the customer and credit card information would be generated automatically by the system and sent to the marketing department for immediate handling.
All analysts need to create an account and log in before they can use the system.
