MIB IIBM CASE STUDY ANSWER SHEETS – What Are Different Stages Of “data Mining”?
Data Mining and Predictive Analytics
- a) Additional acquaintance used by a learning algorithm to facilitate the learning process
- b) A neural network that makes use of a hidden layer
- c) It is a form of automatic learning.
- d) None of these
- Querying of unstructured textual data is referred to as
- a) Information access
- b) Information updation
- c) Information manipulation
- d) Information retrieval
III. A manual component to data mining, consists of preprocessing data to a form acceptable to
- a) Variables
- b) Algorithms
- c) Rules
- d) Processes
- A manual component to data mining, consists t processing data in form of
- a) Discovered processes
- b) Discovered algorithms
- c) Discovered features
- d) Discovered patterns
- Patterns that can be discovered from a given database, can be of
- a) One type only
- b) No specific type
- c) More than one type
- d) Multiple type always
- Analysis tools precompute summaries of very large amounts of data, in order to give
- a) Queries response
- b) Data access
- c) Authorization
- d) Consistency
VII. Data can be store , retrieve and updated in …
- a) SMTOP
- b) OLTP
- c) FTP
- d) OLAP
VIII. Which of the following is a good alternative to the star schema?
- a) snow flake schema
- b) star schema
- c) star snow flake schema
- d) fact constellation
- Background knowledge is…
- a) It is a form of automatic learning.
- b) A neural network that makes use of a hidden layer
- c) The additional acquaintance used by a learning algorithm to facilitate the learning process
- d) None of these
- Which of the following is true for Classification?
- a) A subdivision of a set
- b) A measure of the accuracy
- c) The task of assigning a classification
- d) All of these
- What are data mining techniques? (5)
- What are the applications of data mining? (5)
- Why is data mining important? (5)
- Differentiate Between Data Mining And Data Warehousing? (5)
Section B: Caselets (40 marks)
User-generated content is an indispensable part of today’s industry as every other company needs user data to sell and buy products and provide the best possible support to its users and clients. While user data is important, it needs to be processed to make it relevant for the company. Data mining is the most important tool to process such data and make it relevant and useful.
The decision tree algorithm with the apriori algorithm can be used to support the needs of the client.
To explain this problem, we will turn to smart technology –something that makes our lives easier. Whenever we install any application in our smartphone, we are asked for permission for the installation, but we do not pay too much attention to the information these application require to be installed. In the process, we unknowingly disseminate varied information on maps, massages, contacts, etc. With the help of this information the application, besides collating customer data, also tries to support the users to make their life easier and at the same time makes them dependent on the application in the near future.
Once the user information is gathered, the data is analysed to get the required information so as to give the best information to the algorithm at different times. This type of analysis starts from data pre-processing steps, steps that have already been explained in Chapters 1 and 2. However, for this type of data pre-processing the information gain happens by designing the decision tree at different levels-the depth decision tree or 2-10 level decision tree as well.
Each data gives a valid point of information and these points are used in designing the clusters among different types of data but they are very centric in information as they provide the information of different users according to same contents. The frequency of the matching data is processed by means of decision tree under info gain and Apriori.
It is a common experience nowadays for different applications to recommend the same item for buying from different applications or portals, Users are also able to exercise their choices when it comes to reading the news by selecting the content that is more liked. Through their preferences, they provide the application information about the cognitive behavior of users. This allows prediction of the way a particular4 consumer behaves and recommendations are accordingly tweaked. Most studies of systems or online reviews so far have used only numeric information about sellers or products to examine their economic impact. The understanding that text matters has not been fully realized in electronic markets or in online communities. Insights derived from text mining of user-generated feedback can thus provide substantial benefits to businesses looking for competitive advantages.
Let us summarise some of the chief benefits utiling user-centric data:
It saves money: Since the users themselves provide relevant content for prediction and subsequent recommendations, users data need not be bought and efficiency in terms of time and costs in increased.
It provides variety: By using the user data, the customer can be apprised of various new features or upgrades to the existing product. Further, the user gets to know about the discounts being offered and can avail the support extended to the end user.
It offers a voice to the user: The company is in a position to offer individual customers different products as per individual preferences and a user can provide any specific information of the item he /she wants to use
These benefits of user-centric data should be firmly kept in mind to make such data more predictive and relevant in our fast-paced technological era.
- What do you understand by user generated content? (10)
- Do you really think user generated content is effective? (10)
Big data is the collection and cross-referencing of large numbers and varieties of data sets that allows organizations to identify patterns and categories of cardholders through a multitude of attributes and variables. Every time customers use their cards, big data suggests the products that can be offered to the customers. These days many credit card users receive calls from different companies offering them new credit cards as per their needs and expenses on the existing cards. This information is gathered on the basis of available data provided by vendors.
There are quite a few option available to customers to choose from. Sometimes customers even switch their existing credit card companies. But competition may not always work in the best interests of consumers. It also involves bank’s profit. Competition may also be focused on particular features of credit cards that may not represent long-term value or sustainability.
Those paying interest on balances may be paying more than they realize or expect. Some consumers use up their credit limits quickly or repeatedly make minimum payments without considering how they will repay their credit card debt. A proportion of consumers may also be over-borrowing and taking on too much debt, and there are signs that some issuers may profit more from higher risk borrowers (by which we mean customers at greater risk of credit default).
With the launch of this credit card market study, we intend to build up a detailed picture of the market and assess the potential identified issues. We plan to focus on credit card services offered to retail consumers by credit card providers, including banks, mono-line issuers and their affinity and co-brand partners.
While mass marketing continues to dominate most retailers’ advertising budgets, one-to-one marketing is growing rapidly too. In this case study, you will learn how to improve performance by communicating directly with customers and delighting them with relevant offers. Personalised communication is becoming a norm. Shoppers now expect retailers to provide them with product information and promotional offers that match their needs and desires. They count on you to know their likes, dislikes and preferred communication method-mobile device, email or print media.
On the surface, generating customer-specific offers and communications seems like an unnerving task for many retailers, but like many business problems, when broken into manageable pieces, each process step or analytical procedure is attainable. First, let’s assume you have assembled promotions that you intend to extend as a group of offers (commonly called “offer bank”) to individual customers. Each offer should have a business goal or objective, such as:
Category void for cross or up-selling of a particular product or product group
Basket builder to increase the customer’s basket size
Trip builder to create an additional trip or visit to the store or an additional e-commerce session
Reward to offer an incentive to loyal customers
- How Big data used in this case study- Define? (20)
Section C: Applied Theory (30 marks)
- What Are Olap And Oltp? (15)
- What Are Different Stages Of “data Mining”? (15)