INTELLIGENT TUTORING SYSTEM BASED ON FUZZY LOGIC FOR ADAPTIVE LEARNING OF STATISTICS SURVEY

• Main Author: EKA SEPTIYONO NIM: 23216020, ARIES
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/25758
• Institution: Institut Teknologi Bandung

<p align="justify">Survey training in Statistics Indonesia (BPS) is one of the important stages in conducting the survey, because with the good training, the survey can produce good data quality. Nowadays, many routine statistical survey trainings are eliminated due to budget constraints. One solution to continue training with low cost is selflearning on distance learning (E-Learning). <br /> <br /> <br /> <br /> This research has adopted DRM (Design Research Methodology) which is composed four stages of research: 1) Research Clarification; 2) Descriptive Study I; 3) Prescriptive Study; and 4) Descriptive Study II. At the Prescriptive Study, an E-Learning system that is adaptive to the user's knowledge level has been designed. The system is built on one of the statistical survey activities at BPS, i.e. Annual Manufacturing Survey (SIBS Tahunan). The system is called PASTI (Adaptive Learning for Annual Manufacturing Survey). PASTI is made by combining Intelligent Tutoring System technology, adaptive learning technology, stereotype <br /> <br /> <br /> model, overlay model, fuzzy logic with Mamdani methods and weighted average. <br /> <br /> <br /> <br /> PASTI evaluation uses the first two levels of Kirkpatrick's evaluation model, i.e. evaluation of reactions and evaluation of learning. Evaluation of reactions is measured uses TAM (Technology Acceptance Model), where the proposed system can be accepted by the users and positively impact on the implementation into the daily work. Evaluation of learning uses statistical tests, i.e. paired t-test and Welch t-test. Results for paired t-tests indicated that the system significantly increased users's knowledge by 23%, and results of Welch t-test indicated that learning with the PASTI system was better than conventional learning system.<p align="justify">