IT in Society
Digital currencies (Electronic/Cryptocurrency/CBDC/Blockchain), Data Mining & CRISP-DM (6 stages), Social networking impacts, Impact of IT across sectors, Technology Enhanced Learning.
Danh sách bài học
29 bài📖 Nội dung lý thuyết — T12: IT in Society
Xem bài học chi tiết trong tab Bài học phía trên. Mỗi bài giảng bao gồm:
- ✅ Định nghĩa thuật ngữ chuẩn Cambridge
- ✅ Bảng so sánh & ví dụ minh hoạ
- ✅ Exam tips — cách trả lời câu hỏi Paper 1/3
- ✅ Các lỗi thường gặp cần tránh
- ✅ Sơ đồ & diagram khi cần thiết
Kho đề thi Cambridge theo chương này. Chọn đề để luyện tập có đếm giờ.
Chưa có tài liệu. Giáo viên sẽ bổ sung.
CRISP-DM Data Mining — 6 Stages with Bank Fraud Example
📊 CRISP-DM — 6 Stages
- Business Understanding — Define problem, success metrics, legal constraints with the client
- Data Understanding — Collect historical data; explore variables, quality issues, patterns
- Data Preparation — Clean (remove duplicates, handle missing values); normalise; engineer features (e.g. transaction velocity, distance from home); create balanced training/test sets (SMOTE if class imbalance)
- Modelling — Train ML algorithms (decision trees, random forests, neural networks); tune hyperparameters
- Evaluation — Test on unseen data: accuracy, precision, recall, F1-score, ROC-AUC; assess false positive rate
- Deployment — Integrate into live system; monitor; schedule regular retraining as fraud patterns evolve
⚠️ Bank fraud tip: Fraud events are RARE (class imbalance) — mention oversampling (SMOTE) in Data Preparation for high-mark answers. False positives (blocking legitimate transactions) erode customer trust — precision matters as much as recall.
Mnemonic học thuộc: Electronic payment systems
Cornell Notes: Cryptocurrency overview
Kỹ thuật Feynman: Bitcoin and blockchain
Kết nối với thực tế: CRISP-DM stage 1 Business Understanding
Teach Back (dạy lại): CRISP-DM stage 2 Data Understanding
Active Recall: CRISP-DM stage 6 Deployment
Peer Teaching: Social networking impacts
Ví dụ cụ thể: Data mining techniques
Blockchain — Complete Answer Framework for Paper 3
🔗 Blockchain — All Angles Tested
Describe how blockchain stores data (4 marks):
- Transactions grouped into a block (sender, receiver, amount, timestamp)
- Each block stores the cryptographic hash of the previous block → forms the chain
- Each block has its own hash — tampering changes the hash and invalidates all later blocks
- Stored as a distributed ledger — identical copies across many nodes; consensus mechanism required to add new blocks
Drawbacks (1 mark each — must be described, not just named):
- High energy consumption — Proof of Work consensus requires enormous computational power
- Storage and scalability — entire chain must be stored by all nodes; grows indefinitely
- Transactions irreversible — errors cannot be corrected once confirmed
- GDPR conflict — right to erasure is impossible on an immutable blockchain
Smart contract (4 marks):
Self-executing program stored on blockchain → defines trigger conditions → when conditions met (via oracle) → automatically executes action without human intervention → result permanently recorded.