IndoBloom-AI: An Indonesian AI Curriculum Corpus with Bloom's Taxonomy Annotations

Description: The IndoBloom-AI corpus is a curated collection of 953 learning outcomes extracted from 71 AI-related courses across 21 Indonesian higher education institutions. Each outcome is annotated with Bloom's Taxonomy cognitive levels (L2 Understand, L3 Apply, L4 Analyze, L5 Evaluate, L6 Create) and classified into one of three language categories: Indonesian-only, English-only, or code-mixed Indonesian-English text. This dataset supports research on curriculum-industry alignment, automated cognitive level classification, and code-mixed language processing in educational contexts. The corpus includes bilingual verb anchor lexicons, inter-annotator agreement statistics, and comprehensive metadata for reproducibility. Key Features: 953 annotated learning outcomes 71 courses from 21 institutions across 9 Indonesian regions 5-level Bloom's Taxonomy annotations with Fleiss' κ = 0.856 (validators) Trilingual language stratification: Indonesian-only (50.1%), code-mixed (27.5%), English-only (22.5%) Curated 277-term AI/CS technical terminology lexicon Anonymized institution identifiers (INST-01 to INST-21)