Derican/word2vec-edu-design

Word2Vec of Education Design for General Education Categorization

This is a Word2Vec model trained on a corpus of 836,293 research papers focused on education, design, and performance. The vocabulary was constructed from the papers' titles and abstracts and enriched with bigrams and trigrams of educational terms.

The model captures semantic relationships between concepts and terminology in educational science and practice. It is particularly suitable for tasks such as concept similarity analysis, clustering, and exploring relationships between educational topics.

Corpus and Training Details

• Corpus: 836,293 research papers in English, covering education, design, and performance.
• Vocabulary: Includes unigrams, as well as domain-specific bigrams and trigrams
• Model type: Word2Vec
• Training parameters: For full details, please refer to the original paper.

Intended Use

This model can assist researchers and practitioners in:

• Clustering educational concepts or documents
• Analyzing semantic similarity between educational terms
• Exploring concept relationships in educational research

Limitations

• The model is trained only on English-language papers and may not generalize to other languages.
• It reflects the terminology and focus of the training corpus; domain-specific nuances outside this corpus may not be captured.

Read More: https://arxiv.org/abs/2510.09183