Scholarlink

<p>Database programs like biological science, image extraction, etc., increase the significance of structured data. This work designs a new clustering-based unstructured data analysis model using a deep learning algorithm. Initially, the unstructured data is garnered from online databases, and then it undergoes pre-processing using punctuation removal, “stop word removal, stemming and tokenization operations” to improve the quality of data. Further, the pre-processed data is given to the Layer Improved Transformer Network (LITN) for the retrieval of relevant features from the unstructured data. The retrieved features are subjected to Deep Learning-based Adaptive Clustering (DLAC), where the Deep Neural Network (DNN) is utilized to enhance the clustering performance. Here, the parameters of DNN are optimized utilizing the developed Fitness-based Wild Geese Migration with Cuttlefish Algorithm (FWGMCA). The experiments are finally performed for this designed approach and the implementation results are compared with the existing clustering approaches for ensuring efficiency. From the experiments, the designed FWGMCA-DLAC achieves 94.73% Dunn Index, 94.61% Jaccard Index, 89.53% Silhouette, and 94.56% Hopkins values for the second dataset, which are higher than the traditional techniques. Thus, it is confirmed that the implemented FWGMCA-DLAC highly improves the clustering performance than the existing approaches.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Unstructured data clustering based on layer improved transformer features connected with deep neural network-based adaptive clustering

  • 期刊论文

  • 发表日期：

    2025年9月19日

查看全文
  • Bankapalli Jyothi,
  • L. Sumalatha,
  • Suneetha Eluri

摘要

Database programs like biological science, image extraction, etc., increase the significance of structured data. This work designs a new clustering-based unstructured data analysis model using a deep learning algorithm. Initially, the unstructured data is garnered from online databases, and then it undergoes pre-processing using punctuation removal, “stop word removal, stemming and tokenization operations” to improve the quality of data. Further, the pre-processed data is given to the Layer Improved Transformer Network (LITN) for the retrieval of relevant features from the unstructured data. The retrieved features are subjected to Deep Learning-based Adaptive Clustering (DLAC), where the Deep Neural Network (DNN) is utilized to enhance the clustering performance. Here, the parameters of DNN are optimized utilizing the developed Fitness-based Wild Geese Migration with Cuttlefish Algorithm (FWGMCA). The experiments are finally performed for this designed approach and the implementation results are compared with the existing clustering approaches for ensuring efficiency. From the experiments, the designed FWGMCA-DLAC achieves 94.73% Dunn Index, 94.61% Jaccard Index, 89.53% Silhouette, and 94.56% Hopkins values for the second dataset, which are higher than the traditional techniques. Thus, it is confirmed that the implemented FWGMCA-DLAC highly improves the clustering performance than the existing approaches.