Free unit-wise study notes on testing, results and validation for Major Project, Semester 8 of B.Tech — Computer Science & Engineering — key concepts, examples, important questions and a revision checklist for semester exams.
The crucial phase of proving that the developed system actually solves the problem stated in Unit 1 through rigorous testing methodologies and data validation.
Notebook — 4 pages
Page 1
Wink Notes
B.Tech CSE — 8th Semester
Major Project
— Unit - 4 —
1. The Importance of Testing
A project that 'works on my machine' is not a completed project. Testing is the empirical proof that the software is robust, secure, and meets the requirements.
⇒1.1 Levels of Software Testing
Unit Testing: Testing individual functions or classes in isolation to ensure they return the correct outputs for given inputs.
Integration Testing: Testing the connection between two or more modules (e.g., does the API correctly write to the Database?).
System Testing: Testing the entire compiled application end-to-end.
User Acceptance Testing (UAT): Having actual end-users (or your guide) use the software to verify it solves the business problem.
Page 2
Wink Notes
B.Tech CSE — 8th Semester
Major Project
— Unit - 4 —
2. Performance and Security Validation
Beyond functional correctness, a Major Project must be evaluated for non-functional requirements.
Testing Type
Tool Examples
What it Measures
Load Testing
JMeter, Locust
How the system behaves when 1000 concurrent users hit the API.
Security Testing
OWASP ZAP, SonarQube
Checking for SQL injection, Cross-Site Scripting (XSS), and exposed API keys.
Accessibility Testing
Lighthouse
Ensuring the UI is usable by people with disabilities (screen readers, color contrast).
Page 3
Wink Notes
B.Tech CSE — 8th Semester
Major Project
— Unit - 4 —
3. Results Evaluation (For ML/Data Projects)
If your project involves Machine Learning, 'testing' means evaluating the model against a holdout dataset.
Precision and Recall: Crucial for imbalanced datasets where simple 'Accuracy' is misleading (e.g., cancer detection).
F1 Score: The harmonic mean of Precision and Recall.
You must generate graphs (Loss Curves, ROC-AUC curves) to visually prove in your final report that the model learned properly without overfitting.
Page 4
Wink Notes
B.Tech CSE — 8th Semester
Major Project
— Unit - 4 —
4. Unit Summary and Evaluation Focus
Unit 4 corresponds to the 'Pre-Final Review'.
Edge Cases: Evaluators will actively try to break your system during the demo (e.g., entering letters into a phone number field, clicking submit 10 times rapidly). If the app crashes, you fail the testing review. It must handle errors gracefully.
Quantitative Proof: You cannot just say 'The system is fast.' You must present a slide saying 'The API responds in 120ms under a load of 50 concurrent requests.'
Comparison to Baseline: If you built a new algorithm, you must show a graph comparing its performance against the existing algorithms you mentioned in your Unit 1 Literature Survey.
