Keyword Extraction And Topic Clustering For Education Big Data

Keyword extraction and topic clustering are critical to education big data research.The thesis investigates keyword extraction and topic clustering in the field of education big data using academic paper abstracts.The research content of this thesis is as follows.(1)A semantic keyword extraction algorithm is proposed,which incorporates domain-specific grammar rules and enhanced features derived from adjectives.The algorithm can extract keywords from academic paper abstracts to facilitate topic clustering.The proposed algorithm is validated and evaluated on a dataset of 1028 academic papers from the field of education big data.Experimental results show that the precision of the keyword extractor is 76.8%,outperforming the unsupervised learning-based keyword extractor YAKE and the supervised learning-based keyword extractor LFCN.(2)A topic clustering method that combines autoencoder and k-means algorithm is proposed.Experiments show that Silhouette Coefficient based on the k-means clustering on the public dataset is only 0.0374,whereas the Silhouette Coefficient achieved by the autoencoder combined with the k-means clustering method can reach 0.129.The results indicate that the proposed topic clustering method,which leverages autoencoder and k-means,outperforms traditional k-means methods.(3)By applying the proposed method to educational big data,12 meaningful clusters are identified.Grouping academic papers based on these topics can provide valuable support for researchers in this field.