# UAS — Artificial Intelligence (CSA) (2024)

> Final Exam · Odd Semester 2024/2025 · Universitas Gadjah Mada

| Field | Value |
|---|---|
| **Course** | Artificial Intelligence (CSA) |
| **Exam type** | Final Exam (UAS) |
| **Semester** | Odd Semester 2024/2025 |
| **Date** | Monday, 2 December 2024 |
| **Time** | 100 minutes |
| **Lecturer** | Dzikri Rahadian Fudholi, S.Kom., M.Comp. |
| **Exam mode** | Open one A4/F4 Sheet |

**Institution:** Program Studi S1 Ilmu Komputer, Fakultas Matematika dan Ilmu Pengetahuan Alam, Universitas Gadjah Mada

*Follow the examiner rules and good luck!*

---

**1. PLO 2 & 3 - CO 4 - Able to explain Expert System Architecture (7%)**

- **A.** Explain the role of the knowledge base, inference engine, and user interface in an expert system! Illustrate your explanation with an example of how these components interact in a specific domain (e.g., medical diagnosis or financial planning).
- **B.** Discuss the challenges in maintaining the knowledge base over time.

**2. PLO 2 & 3 - CO 5 - Able to explain Natural Language Processing system (8%)**

- **A.** Discuss the importance of text preprocessing in NLP tasks.
- **B.** Describe the steps involved in creating a preprocessing pipeline for an emotion recognition system, addressing challenges such as handling noisy data and tokenization in morphologically rich languages.
- **C.** Explain the traditional word representations of TF-IDF! Discuss why this method will not perform well.

**3. PLO 4 & 5 - CO 6 - Able to explain Pattern Recognition and Machine Learning Methods (10%)**

An e-commerce company wants to develop a system to classify customers into "High-Risk" or "Low-Risk" groups for fraudulent transactions based on numerical features. The dataset includes transaction amounts, frequency of purchases, account age, refund rates, and other behavioural metrics. The goal is to accurately classify customers while minimizing false positives (flagging legitimate customers as high-risk) and false negatives (missing fraudulent customers).

**Task:**
Write an essay/proposal preferably in points, addressing the following:

1. **Data Exploration and Preprocessing:**
   - Discuss how to handle missing values, outliers, and skewed distributions in numerical features.
   - Propose techniques for feature selection or dimensionality reduction to enhance model interpretability and performance.
2. **Model Selection and Training:**
   - Compare classification models for this task.
   - Explain how you would tune hyperparameters to optimize classification performance.
3. **Evaluation and Class Imbalance:**
   - Identify appropriate evaluation metrics and justify your choices.
   - Discuss methods to handle imbalanced classes.

---

>>> Best of Luck to you ^-^v <<<

| | |
|---|---|
| Telah direview pada tanggal: …………….. | Tanda tangan Reviewer: |
| Nama Reviewer: ………………………….. | |

## Transcription notes

- **Source:** `UAS AI.pdf` (2 pages, image-only scan; rendered and read visually).
- **This is the CSA-class paper.** It is textually **identical** to the CSB paper transcribed in `UAS - AI - 2025_1.md` (from `UAS - AI - 2025_1.jpg` + `_2.jpg`) except for the class label: **CSA vs CSB**. Same date, lecturer, questions and weightings.
- Because the two are duplicates apart from the class code, they are kept as separate files (they are separate source documents) but should be treated as one exam when studying.