(SEM VI) THEORY EXAMINATION 2022-23 DATA ANALYTICS

DATA ANALYTICS – KIT-601

Section-wise Important Questions & Ready Answers

SECTION A

(Attempt all questions – 2 marks each)

(a) Main Characteristics of Big Data

Big Data is characterized by Volume (huge amount of data), Velocity (high speed of data generation), Variety (different data types), Veracity (data uncertainty), and Value (usefulness of data). These characteristics differentiate Big Data from traditional data systems.

(b) Role of Analytical Tools in Big Data

Analytical tools help in collecting, cleaning, processing, analyzing, and visualizing large datasets. They transform raw data into meaningful insights that support decision-making, prediction, and optimization.

(c) Purposes of Regression Analysis

Regression analysis is used to study the relationship between dependent and independent variables. It helps in prediction, trend analysis, forecasting future values, and understanding the impact of variables.

(d) Fuzzy Qualitative Model

A fuzzy qualitative model represents uncertain or imprecise information using fuzzy logic instead of exact numerical values. It is useful when human reasoning or linguistic variables are involved.

(e) Association Rule

An association rule identifies relationships between items in a dataset. It is expressed in the form X → Y, meaning if X occurs, Y is likely to occur.

(f) Benefits of Analytic Sandbox

An analytic sandbox provides an isolated environment where analysts can explore data, test models, and perform experiments without affecting production systems.

(g) Data Stream Management System (DSMS)

A DSMS processes continuous, real-time data streams instead of static datasets. It supports querying, filtering, and aggregation of streaming data.

(h) Response Modeling

Response modeling predicts how users or customers will respond to certain actions such as advertisements, offers, or campaigns using historical data.

(i) Benefits of Visual Data Exploration

Visual data exploration helps users identify patterns, trends, and anomalies quickly. It improves understanding, reduces analysis time, and supports better decisions.

(j) Main Goals of Hadoop

Hadoop aims to provide scalable storage, distributed processing, fault tolerance, and cost-effective handling of Big Data using commodity hardware.

SECTION B

(Attempt any three – 10 marks each)

2(a) Analysis vs Reporting in Data Analytics

Reporting focuses on presenting historical data in predefined formats such as dashboards and reports. Analysis goes beyond reporting by exploring data, identifying patterns, predicting outcomes, and supporting strategic decisions. Reporting answers what happened, while analytics answers why it happened and what will happen next.

2(b) Neural Network and Its Use in Analytics

A neural network is a computational model inspired by the human brain. It consists of layers of interconnected neurons that process data. In analytics, neural networks are used for classification, prediction, image recognition, fraud detection, and pattern recognition due to their ability to learn complex relationships.

2(c) Apriori Association Rule Mining Algorithm

The Apriori algorithm identifies frequent itemsets by using the principle that all subsets of a frequent itemset must also be frequent. It generates candidate itemsets, calculates their support, and eliminates infrequent ones iteratively until all strong association rules are found.

2(d) Advantages and Disadvantages of K-Means Clustering

K-Means is simple, fast, and efficient for large datasets. However, it requires predefined number of clusters, is sensitive to initial centroids, and performs poorly with non-spherical or unevenly sized clusters.

2(e) HDFS and Handling of Big Data

HDFS is a distributed file system that stores large datasets across multiple machines. It divides data into blocks, replicates them for fault tolerance, and allows parallel access, making it suitable for Big Data storage.

SECTION C

3(a) Big Data Analytics Life Cycle

The Big Data analytics life cycle includes data discovery, data preparation, model planning, model building, result evaluation, and deployment. Each stage ensures structured handling of data from raw form to actionable insights.
(In exam, a neat labeled diagram is expected.)

3(b) Regression Modeling vs Bayesian Modeling

Regression modeling estimates relationships using fixed parameters, while Bayesian modeling treats parameters as random variables and updates beliefs using prior and posterior probabilities. Bayesian models handle uncertainty better than traditional regression.