{
  "id": "uas-ai",
  "subject": "Kecerdasan Artifisial",
  "exam": "UAS",
  "year": null,
  "variant": null,
  "sources": [
    {
      "file": "UAS/UAS AI.pdf",
      "type": "pdf",
      "part": 0
    }
  ],
  "pageCount": 2,
  "charCount": 2686,
  "pages": [
    {
      "source": "UAS/UAS AI.pdf",
      "page": 1,
      "method": "ocr",
      "text": "1\nMWA PROGRAM STUDI S1 ILMU KOMPUTER\nAO) NI FAKUL VAS MATEMATIKA DAN ILMU PENGETAHUAN ALAM\nNA UNIVERSITAS GADJAH MADA\nSo ———— ee ——\nLinn Exim\nOdd Semester 2024/2028\nCourse Name : Artificial Intelligence (CSA)\nDay, Date : Monday, 2 December 2024\nTime : 100 minutes . ,\nLecturer : Dzikri Rahadian Fudholi, S.Kom., M.Comp,\nExam mode : Open one A4/F4 Sheet\noo\nLollow the examiner rules and good luck!\n1. PLO2& 3- CO 1- Able to explain Papert System Architecture (7%)\nA. Explain the role of the knowledge base, inference engine, and user interface in an\nexpert system! [trate your explanation witli ample of how these components\ninteract in a specific domain (6.1. medical diagnosis or financial planning).\nB. Discuss the Ciallenges in maintaining the knowledye base over time\n2. PLO2& 3-CO5-Able to explain Natural Language Processing system (8%)\nA. Discuss the importance of text preprocessing in NLP 11 ks.\nB. Describe the steps involved in creating a preprocessing pipeline for an emotion\nrecognition system, addressing chollenges such as handling 1eisy data and token: tion in\nmorphologically rich languages.\nC. Explain the traditional word representations 01 1T-IDF! Discuss why thi method will\nnot perform well.\n3. PLO 4 & 5- CO 6- Able to explain Pattern Recognition and Machine Learning\nMethods (10 %)\nAn e-commerce company wants to develop a system to classify customers into \"tigh-\nRisk\" or \"Low-Risk\" groups for fraudulent transactions based on numerical features: The\nSEX or LOWE ) EN ee\ndatascl includes lransaction ame, nis. frequency of purchases, account age, refund rates,\nand other behavioural mefrics. The goal is to accurately classify customers while\nminimizing false positives (flagging legitimate customers as high-risk) and ful .c negatives\n(missing fraudulent customers).\nTask:\nWrite an essay/proposal preicrably in points, addressing tie following:\n. Tanda tangan Reviewer:\nTelah direview pada tanggal: conn.\nNama Reviewer: .o........o.... 1 cos. 10010eneaaan |"
    },
    {
      "source": "UAS/UAS AI.pdf",
      "page": 2,
      "method": "ocr",
      "text": "a we 2 Be ee al\n1. Data Exploration and Preprocessing:\no Discus how to handle missing values, outliers, and skewed distributions in\nnumerical features.\no Propose techniques for feature selection or dimensionality reduction to\nenhance model interpretability and performance.\n2. Model Selection and Training:\no Compare classification models for this task.\no Explain how you would tune hyperparameters to optimize classification\nperformance.\n3. Evaluation and Class Imbalance:\no Identify appropriate evaluation metrics and ju-tify your choices.\no Discuss method: to handle imbalanced classes.\n>>> Best of Luck to you \"vy <<<\nTanda tangan Reviewer:\n| Telah direview pada tang 9.\n| Nama Reviewer: meet ER\n| |"
    }
  ],
  "text": "1\nMWA PROGRAM STUDI S1 ILMU KOMPUTER\nAO) NI FAKUL VAS MATEMATIKA DAN ILMU PENGETAHUAN ALAM\nNA UNIVERSITAS GADJAH MADA\nSo ———— ee ——\nLinn Exim\nOdd Semester 2024/2028\nCourse Name : Artificial Intelligence (CSA)\nDay, Date : Monday, 2 December 2024\nTime : 100 minutes . ,\nLecturer : Dzikri Rahadian Fudholi, S.Kom., M.Comp,\nExam mode : Open one A4/F4 Sheet\noo\nLollow the examiner rules and good luck!\n1. PLO2& 3- CO 1- Able to explain Papert System Architecture (7%)\nA. Explain the role of the knowledge base, inference engine, and user interface in an\nexpert system! [trate your explanation witli ample of how these components\ninteract in a specific domain (6.1. medical diagnosis or financial planning).\nB. Discuss the Ciallenges in maintaining the knowledye base over time\n2. PLO2& 3-CO5-Able to explain Natural Language Processing system (8%)\nA. Discuss the importance of text preprocessing in NLP 11 ks.\nB. Describe the steps involved in creating a preprocessing pipeline for an emotion\nrecognition system, addressing chollenges such as handling 1eisy data and token: tion in\nmorphologically rich languages.\nC. Explain the traditional word representations 01 1T-IDF! Discuss why thi method will\nnot perform well.\n3. PLO 4 & 5- CO 6- Able to explain Pattern Recognition and Machine Learning\nMethods (10 %)\nAn e-commerce company wants to develop a system to classify customers into \"tigh-\nRisk\" or \"Low-Risk\" groups for fraudulent transactions based on numerical features: The\nSEX or LOWE ) EN ee\ndatascl includes lransaction ame, nis. frequency of purchases, account age, refund rates,\nand other behavioural mefrics. The goal is to accurately classify customers while\nminimizing false positives (flagging legitimate customers as high-risk) and ful .c negatives\n(missing fraudulent customers).\nTask:\nWrite an essay/proposal preicrably in points, addressing tie following:\n. Tanda tangan Reviewer:\nTelah direview pada tanggal: conn.\nNama Reviewer: .o........o.... 1 cos. 10010eneaaan |\n\na we 2 Be ee al\n1. Data Exploration and Preprocessing:\no Discus how to handle missing values, outliers, and skewed distributions in\nnumerical features.\no Propose techniques for feature selection or dimensionality reduction to\nenhance model interpretability and performance.\n2. Model Selection and Training:\no Compare classification models for this task.\no Explain how you would tune hyperparameters to optimize classification\nperformance.\n3. Evaluation and Class Imbalance:\no Identify appropriate evaluation metrics and ju-tify your choices.\no Discuss method: to handle imbalanced classes.\n>>> Best of Luck to you \"vy <<<\nTanda tangan Reviewer:\n| Telah direview pada tang 9.\n| Nama Reviewer: meet ER\n| |"
}