The automated construction of educational knowledge graphs from programming tutorial websites faces significant challenges due to the heterogeneous nature of web content structures and the complex relationships between code snippets and their contextual information. While existing approaches rely heavily on either manual rule crafting or direct large language model (LLM) processing, both methods struggle with cross-platform scalability and extraction accuracy. We present a novel distribution-guided extraction approach that systematically analyzes HTML structural patterns through strategic sampling before deploying LLM-based extraction scripts. This approach significantly reduces manual engineering effort while maintaining high accuracy across diverse platforms. Our method achieves a substantial improvement in code snippet extraction accuracy (from 0.15–0.68 to 0.54–0.91) across major programming tutorial websites including w3schools, MDN, and runoob. The extracted knowledge is integrated into a comprehensive educational graph that powers a question-answering system, achieving faithfulness scores of 0.93 and answer correctness of 0.90. These results demonstrate the effectiveness of combining statistical pattern analysis with LLM-driven extraction for building robust, cross-platform educational knowledge systems.

Users often use tags to annotate and categorize web content. A folksonomy is a system of classification derived from the practice and method of collaboratively creating and managing tags. The most significant feature of a folksonomy is that it directly reflects the vocabulary of users. This feature is very useful in tag-based content searching and user browsing. Based on mutual-overlapping measurement of tag's instance sets, an ontology learning algorithm to construct concept hierarchy from folksonomy is proposed. A case study of datasets from a famous Chinese e-business website taobao is carried out. The precision, valid, recall and F-measure rates of the constructed concept hierarchy are 54%, 84%, 100% and 70% respectively. The experimental results on real world datasets show that the proposed method is feasible.

Because of the surprisingly increasing volume and complicated nature of remote sensing images (RSIs), one of the main obstacles to realize efficient retrieval of the RSIs is the lack of effective sharing technologies and semantic description methods. In this letter, we design and implement a prototype grid system named RSIsGrid for semantic-based RSIs retrieval using ontology and grid technologies. In order to verify the semantic-based method, measures such as the recall, precision, and query time are used. Test results have indicated that the semantic-based method can promote the query performance of the RSIs.