Summary: Business Intelligence & Business Analytics

Read the summary and the most important questions on Business Intelligence & Business Analytics

• 1 BIBA

• 1.1.2 Introduction to Databases

  This is a preview. There are 4 more flashcards available for chapter 1.1.2
  Show more cards here

• What is a database and what does is consist of?

  • 
    Definition database: “A collection of related tables, designed, maintained and utilized by multiple users, with software to update & query the data”
    Database system consists of 
    
    • 1. Data (the database) 
      
    • 2. Software 
      
    • 3. Hardware 
      
    • 4. Users 
      

• What is a database management system (DBMS) and how is it used?

  Database management system (DBMS) is the software than controls the data 
  

  • Oracle, DB2, MS Access, MS SQL Server (Azure) 
    
  • MySQL (open source)
    

  Manipulation of data using a query language 
  

  • E.g. SQL (Structured Query Language) 

  
  

• What are the database terminology?

  • Database 
  • Table = structured list of data of a specific type, divided by columns and rows 
  • Record/Tuple 
  • Field/Attribute 
  • Domain

• 1.1.3 Relational database

  This is a preview. There are 3 more flashcards available for chapter 1.1.3
  Show more cards here

• What is a primary key and foreign key?

  PK:  Field(s) that uniquely identifies each record in a table 
  

  • Bno (Book), Rno (Reader), Bno + Rno + Loan date (Loan) 
    
  • Null value (=no data entry) not allowed for PK
    

  FK: Attribute whose values match the primary key values in the related (parent) table 
  

  • E.g.: Vendor sells Products 
    
  • Precisely 1 vendor per product 
    
  • Conversely, a vendor might sell multiple products 
    
  • FK = ‘Vendor_Code’ [Product table]
    

• What are 4 trends in the database world?

  • Trend 1: From disk-based to In-memory databases (e.g. SAP HANA)
    

  • Trend 2: From on-premise db to cloud db (e.g. MS Azure, Google cloud sql) 
    

  • db use becomes an operating expense instead of capital expense
    
  • Trend 3: No (not only) SQL databases
    

  • Hypothesis: For analytics, relational database are dominant
    
  • NoSQL Databases: Key-value, Document, Graph, Wide-column
    
  • Trend 4: Alternative data representations
    

  • For storing document-oriented files with hierarchies; use XQuery
    
  
  

• 1.2.3 Data warehouse architectures

  This is a preview. There are 5 more flashcards available for chapter 1.2.3
  Show more cards here

• Which DW development approaches are there and which is the best?

  • Data mart approach (bottom-up)
    
  • DW = a collection of data marts
    
  • Dimensional modeling
    
  • Consistency achieved by conformed dimensions
    
  • E.g.: Independent data marts, Bus, ‘Canned data warehouse’
    
  • Enterprise dw approach (top-down)
    
  • DW = one integrated database
    
  • Entity-relationship modeling
    
  • E.g. Hub & spoke: EDW + dep. data marts, Federated DW
    
  • Which approach is best?
    
  • There is no one-size-fits-all strategy to DW, depending on: management’s information needs, inf. interdependence between organizational units, …
    

• 1.6.2 Naïve Bayes

  This is a preview. There are 3 more flashcards available for chapter 1.6.2
  Show more cards here

• Why is laplace smoothing needed?

  From previous slides, the probability of α given class c:
  

  • P( Outlook=“Sunny” | PlayTennis=“Yes” ) = 0
    
  • Problem:
    
  • An attribute value doesn’t occur with every class
    
  • Probability of α given class c becomes 0
    
  • Having a probability zero is problematic, because it wipes out all information in other probabilities
    

• What is laplace smoothing?

  • Laplace Smoothing, or Correction, or Estimator 
    
  • Incorporates a small-sample correction in every probability computation
    
  • Increase the numerator/denominator 
    
  • Thus, no probability will be zero
    

• What are advantages and disadvantages for Naive Bayes?

  • Naive Bayes is Not So Naïve:
    
  • Its beauty is in its simplicity
    
  • Ability to handle categorical variables directly
    
  • Computational efficient
    
  • Good classification performance, especially when the number of predictors is very large
    
  • Negative aspects:
    
  • Requires a very large number of records to obtain good results
    
  • Independence assumption may not hold for some attributes
    

• 1.7.1 Quizes

  This is a preview. There are 3 more flashcards available for chapter 1.7.1
  Show more cards here

• Consider you were given 8 items, i.e., records, with numerical variables X1 & X2 along with a dependent variable y that corresponds to color (blue/red). Plot illustrates data in 2D space Task → use k-Nearest neighbors with Euclidean distance to classify item (1,1)Predict the class of new item (1,1) when using k-Nearest neighbors with Euclidean distance and K=3.1. Class of item (1,1) is red.2. Class of item (1,1) is blue

  Red