THEORY EXAMINATION (SEM–VI) 2016-17 DATABASE MANAGEMENT SYSTEM, DATA MINING & WAREHOUSING

DBMS, DATA MINING & WAREHOUSING – ECS019

B.Tech (SEM VI) | Section-wise Solved Answers

SECTION – A

(2 marks each)

(a) Weak Entity

A weak entity is an entity that cannot be uniquely identified by its own attributes and depends on a strong entity.
Example: Dependent depends on Employee.

(b) Precedence Graph for Serializability

A precedence graph is constructed by representing transactions as nodes and conflicts as directed edges. If the graph has no cycle, the schedule is serializable.

(c) Fragmentation and its Types

Fragmentation divides a database into smaller parts. Types include horizontal, vertical, and hybrid fragmentation to improve performance and locality.

(d) Multivalued Dependency

A multivalued dependency occurs when one attribute determines multiple independent values of another attribute.

(e) Use of Knowledge Base

A knowledge base stores domain knowledge, rules, and patterns to support decision-making in data mining.

(f) Hierarchical vs Network Data Model

Hierarchical model follows a tree structure with one-to-many relationships, while network model supports many-to-many relationships using pointers.

(g) Confidence and Support

Support measures frequency of an itemset.
Confidence measures the reliability of an association rule.

(h) Nominal vs Ordinal Variable

Nominal variables have no order (e.g., gender).
Ordinal variables have a meaningful order (e.g., grades).

(i) Integration in KDD Process

Integration combines data from multiple sources into a unified dataset for mining.

(j) Box Plot

Box plot shows data dispersion using quartiles, median, and outliers.

SECTION – B

(Attempt any five)

(a) Classification and Prediction

Classification assigns class labels to data objects, while prediction estimates numeric values.
Example method: Decision Tree Classification.

(b) Concept Hierarchy & Data Mart

Concept hierarchy organizes data from low to high levels of abstraction.
A data mart is a subset of a data warehouse focused on a specific business area.

(c) Types of OLAP Servers

OLAP servers are classified as MOLAP, ROLAP, and HOLAP, based on storage and processing.

(d) Clustering vs Classification

Clustering groups data without predefined classes, while classification uses labeled data.
Example: K-Means clustering.

(e) Specialization & Generalization (ER Model)

Specialization divides an entity into subclasses, while generalization combines entities into a superclass.
Bank ER Example: Accounts, Loans, Branches.

(f) Normalization & 3NF

Normalization removes redundancy.
A relation is in 3NF if it is in 2NF and no transitive dependency exists.

(g) Non-loss Decomposition & Query Optimization

Non-loss decomposition ensures original relation can be reconstructed.
Query optimization evaluates expressions to minimize cost.

(h) ACID Properties & Shadow Paging

ACID ensures reliable transactions.
Shadow paging maintains database consistency without logs.

SECTION – C

(Attempt any two)

(3) Market Basket Analysis & Apriori

Market Basket Analysis discovers item associations.
Apriori algorithm finds frequent itemsets using support thresholds.

(4) Short Explanations

PAM: Partitioning Around Medoids clustering algorithm
Hierarchical Clustering: Builds clusters using tree structure
Serializability: Ensures correct concurrent execution
Deadlock Handling: Prevents or resolves deadlocks
Log-based Recovery: Uses logs to recover data after failure

(5) SQL Queries

-- i SELECT RollNo, Name FROM Student WHERE Branch='CSE'; -- ii SELECT Name FROM Student WHERE RollNo IN (  SELECT RollNo FROM SP  WHERE ISBN IN (    SELECT ISBN FROM Book WHERE Publisher='ABC'  ) ); -- iii SELECT Title, Author FROM Book WHERE ISBN IN (  SELECT ISBN FROM SP WHERE RollNo='XYZ' ); -- iv SELECT Title FROM Book WHERE ISBN IN (  SELECT ISBN FROM SP WHERE Date_of_issue <= '2011-01-01' ); -- v SELECT * FROM Book WHERE Publisher='ABC';