MCS-221 Data Warehousing and Data Mining: Complete Study Guide for IGNOU MCA 2026-27
If you are an IGNOU MCA student, MCS-221 (Data Warehousing and Data Mining) is one of the most useful courses of Semester 2. The MCS-221 assignment 2026-27 (code MCA_NEW/MCAOL(II)/221/Assign/2026-27) has 10 compulsory questions of 8 marks each, plus 20 marks for the viva voce.
In this guide you get a simple explanation of every question, the submission dates, a free downloadable Word file and viva tips.
MCS-221 Assignment 2026-27 – Free Download
All 10 questions answered in simple language with diagrams and examples.
⬇ Download MCS-221 Assignment (DOCX)download •PDF • Updated October 2026
1. MCS-221 assignment details and last dates
| Item | Detail |
|---|---|
| Course | MCS-221 Data Warehousing and Data Mining |
| Programme | MCA_NEW / MCAOL, Semester 2 |
| Assignment number | MCA_NEW/MCAOL(II)/221/Assign/2026-27 |
| Maximum marks | 100 (80 assignment + 20 viva voce) |
| Weightage | 30% |
| Last date (July 2026 session) | 31 October 2026 |
| Last date (January 2027 session) | 15 April 2027 |
| How to submit | MCA_NEW: to your Study Centre coordinator. MCAOL: on the LMS portal. |
Always confirm the latest dates on ignou.ac.in.
2. Question-wise topics (Q1 to Q10)
| Q. | Topic | Marks |
|---|---|---|
| Q1 | Need and characteristics of a data warehouse; Inmon vs Kimball (university case) | 8 |
| Q2 | Three-tier data warehouse architecture (healthcare case) | 8 |
| Q3 | Dimensional model and star schema (supermarket case) | 8 |
| Q4 | ETL from e-commerce, ERP and CRM; data quality issues | 8 |
| Q5 | OLTP vs OLAP; roll-up, drill-down, slice, dice, pivot | 8 |
| Q6 | Data preprocessing for noisy financial data | 8 |
| Q7 | Association rules, Apriori, support, confidence, lift | 8 |
| Q8 | Decision tree, Naive Bayes, k-NN, SVM; credit-risk prediction | 8 |
| Q9 | Partitioning, hierarchical, density-based clustering; K-means vs DBSCAN | 8 |
| Q10 | Emerging trends: cloud DW, big data, text, web and stream mining | 8 |
3. What is a data warehouse and why is it needed? (Q1, Q2)
A data warehouse is a central store that collects data from many systems, cleans it and keeps its history, so managers can analyse it. In a university, admissions, exams, fees and the LMS each have their own database. A warehouse joins them so you can ask questions like "Do students who pay fees late also score lower?"
Four characteristics (Bill Inmon)
- Subject-oriented: organised by topics such as student or course.
- Integrated: data from different sources is made consistent.
- Time-variant: history is kept, such as 10 years of results.
- Non-volatile: data is added, not overwritten.
Inmon vs Kimball
| Point | Inmon (top-down) | Kimball (bottom-up) |
|---|---|---|
| Start | Central enterprise warehouse | Department data marts |
| Model | Normalised (3NF) | Star schema |
| Speed | Slower first result | Faster first result |
| Integration | Very strong | Through conformed dimensions |
Verdict: for a university joining four systems, Inmon is more suitable. A hybrid (Inmon core + Kimball marts) is also popular.
Three-tier architecture (Q2)
- Bottom tier: data sources, staging area, ETL, warehouse and metadata repository.
- Middle tier: OLAP server (ROLAP, MOLAP or HOLAP).
- Top tier: front-end tools such as reports, dashboards and data-mining tools.
In a hospital, the sources are patient records, lab systems, pharmacy and insurance claims. Patient names are masked during ETL to protect privacy.
4. Star schema design (Q3)
A star schema has one central fact table (the numbers) surrounded by dimension tables (the context). For a supermarket:
- Grain: one row per product on each bill.
- Fact table (Fact_Sales): quantity, sales amount, discount, cost, profit.
- Dimensions: Date, Product, Store, Customer, Promotion.
Why star schema? Queries are fast, the design is simple, business users understand it easily and measures can be added up along any dimension. Use Type-2 slowly changing dimensions to keep old history when a customer changes city.
Want the full solution file?
Q1 to Q10 ready as a Word file. Use it as a reference and write in your own words.
⬇ Download Full Solution Freedownload • PDF • Updated October 2026
5. ETL process (Q4)
ETL = Extract, Transform, Load.
- Extract: pull data from e-commerce, ERP and CRM into a staging area (first a full load, then incremental loads).
- Transform: clean, standardise, remove duplicates, match customers across systems and create surrogate keys.
- Load: load dimension tables first, then fact tables.
| Data problem | Example | Fix |
|---|---|---|
| Missing values | No phone number | Default value or "Unknown" |
| Duplicates | Same customer twice | Fuzzy matching, master key |
| Different formats | 12/03/26 vs 2026-03-12 | Standard date and currency |
| Invalid values | Age = 250 | Validation rules, error table |
| Conflicts | Different address in ERP and CRM | Source priority, latest value wins |
6. OLTP vs OLAP and OLAP operations (Q5)
| Feature | OLTP | OLAP |
|---|---|---|
| Purpose | Daily transactions | Analysis and decisions |
| Users | Clerks, customers | Managers, analysts |
| Data | Current, detailed | Historical, summarised |
| Design | Normalised | Star / snowflake |
| Example | ATM withdrawal | Yearly sales by region |
Five OLAP operations
- Roll-up: city sales to state sales.
- Drill-down: quarter to month to day.
- Slice: fix one dimension (for example only Q1).
- Dice: choose values on two or more dimensions.
- Pivot: rotate rows and columns to get a new view.
7. Data preprocessing (Q6)
Dirty data gives wrong results, so five steps are done before mining:
- Cleaning: fill missing values (mean, median, prediction) and smooth noise using binning or outlier detection.
- Integration: merge sources, match the same customer and remove redundant attributes.
- Transformation: normalise values. Min-max formula:
v' = (v - min) / (max - min). Example: income 30,000 with min 20,000 and max 1,00,000 gives 0.125. - Reduction: remove unneeded attributes, use PCA or sampling.
- Discretization: turn numbers into ranges, such as income into Low, Medium and High.
Binning example: ages 4, 8, 15, 21, 21, 24, 25, 28, 34 go into three bins of three values. Replacing by the bin mean gives 9, 22 and 29.
8. Apriori algorithm with example (Q7)
Association rules find items bought together, written as A ⇒ B.
- Support: how often A and B appear together.
- Confidence: support(A and B) / support(A).
- Lift: confidence / support(B). Above 1 means a real positive link.
Apriori uses one idea: if an itemset is rare, all its supersets are also rare, so they are skipped.
| TID | Items |
|---|---|
| T1 | Bread, Milk |
| T2 | Bread, Diaper, Beer, Eggs |
| T3 | Milk, Diaper, Beer, Cola |
| T4 | Bread, Milk, Diaper, Beer |
| T5 | Bread, Milk, Diaper, Cola |
With minimum support 60% (3 of 5 transactions), the frequent single items are Bread, Milk, Diaper and Beer. The frequent pairs are {Bread, Milk}, {Bread, Diaper}, {Milk, Diaper} and {Diaper, Beer}. No 3-item set reaches support 3, so the algorithm stops.
| Rule | Support | Confidence | Lift |
|---|---|---|---|
| Diaper ⇒ Beer | 60% | 75% | 1.25 |
| Beer ⇒ Diaper | 60% | 100% | 1.25 |
| Bread ⇒ Milk | 60% | 75% | 0.94 |
Bread ⇒ Milk looks strong, but its lift is below 1, so it is not a useful rule. This is why lift matters along with confidence.
9. Classification: which algorithm for credit risk? (Q8)
| Algorithm | Strength | Weakness |
|---|---|---|
| Decision tree | Easy to explain, handles mixed data | Can overfit (use pruning) |
| Naive Bayes | Very fast, gives probabilities | Assumes attributes are independent |
| k-NN | Simple, no training | Slow on big data |
| SVM | High accuracy | Hard to explain, needs tuning |
For credit-risk prediction, a decision tree is the best choice here because banks must explain why a loan was approved or rejected. Handle class imbalance with class weights and judge the model with recall and AUC, not only accuracy.
10. K-means vs DBSCAN (Q9)
- K-means: choose k, assign each point to the nearest centre, recalculate centres and repeat until they stop changing. Fast, but sensitive to outliers and finds only round clusters.
- DBSCAN: groups dense regions using ε (radius) and MinPts. It marks core, border and noise points, and finds clusters of any shape.
| Point | K-means | DBSCAN |
|---|---|---|
| Noise | Pushed into clusters | Marked as noise |
| Shape | Round only | Any shape |
| Need k? | Yes | No (needs ε and MinPts) |
| Best for | Clean data, segmentation | Noisy data, fraud, GPS data |
For noisy data, DBSCAN is the better choice.
11. Emerging trends (Q10)
- Cloud data warehouses (BigQuery, Snowflake, Redshift): scalable and pay-as-you-go.
- Big data analytics: handles large, fast and varied data using Hadoop and Spark.
- Text mining: finds insights in reviews, emails and complaints.
- Web mining: studies page content, links and user clicks.
- Data stream mining: analyses live data for instant fraud alerts.
Together they take business intelligence from "what happened" to "what will happen".
Ready for the viva?
Download the file, practise the Apriori and K-means examples, and revise before the viva.
⬇ Download MCS-221 Solutiondownload • PDF • Updated October 2026
12. Viva voce tips
- Know your own answers and be able to explain every diagram.
- Practise the Apriori calculation and the K-means example by hand.
- Be ready to say why you chose a star schema and why a decision tree for credit risk.
- Carry your assignment copy and ID card.
13. Frequently asked questions
What is the last date to submit the MCS-221 assignment 2026-27?
31 October 2026 for the July 2026 session and 15 April 2027 for the January 2027 session.
Is the viva voce compulsory for MCS-221?
Yes. If you submit the assignment but skip the viva, it is marked zero.
How many questions are in the MCS-221 assignment?
There are 10 compulsory questions of 8 marks each, plus 20 marks for the viva.
Where do MCAOL students submit the assignment?
On the IGNOU LMS portal. MCA_NEW students submit at their Study Centre.
Can I copy the solution as it is?
No. Write it in your own words. You must explain it in the viva.
Was this guide helpful?
Join our Facebook Page for MCS-218, MCS-219 and MCS-220 updates and last-date alerts.
Join Facebook PageRelated guides on etutor12
- MCS-218 Solved Assignment 2026-27 (Data Communication and Computer Networks)
- IGNOU MCS-220 solved assignment 2026-27
- All IGNOU MCA solved assignments
Disclaimer: etutor12 is an independent study blog and is not affiliated with IGNOU. This file is for reference only. Always confirm dates and rules on ignou.ac.in. Last updated: October 1, 2026.