dc.description.abstract | This study combines Term Frequency-Inverse Document Frequency technique with compatibility and applies it to the “Regulations of National Central University and Extensions of Off-campus Regulations” and establishes them on the cloud platform for tax classification.
Term Frequency-Inverse Document Frequency technique can only present one type of measurement and quantitative method and is not capable of presenting diverse selection. Therefore, through the combination of compatibility, Cosine Similarity, Hierarchical Clustering and other techniques, a regulation can produce different results in different compatibility. A wide range of selection can be produced through classification, helping users to find the proper regulations which is related.
keyword:text mining、TF-IDF、Cosine Similarity、Hierarchical Clustering
This study combines Term Frequency-Inverse Document Frequency technique with compatibility and applies it to the “Regulations of National Central University and Extensions of Off-campus Regulations” and establishes them on the cloud platform for tax classification.
Term Frequency-Inverse Document Frequency technique can only present one type of measurement and quantitative method and is not capable of presenting diverse selection. Therefore, through the combination of compatibility, Cosine Similarity, Hierarchical Clustering and other techniques, a regulation can produce different results in different compatibility. A wide range of selection can be produced through classification, helping users to find the proper regulations which is related.
keyword:text mining、TF-IDF、Cosine Similarity、Hierarchical Clustering
This study combines Term Frequency-Inverse Document Frequency technique with compatibility and applies it to the “Regulations of National Central University and Extensions of Off-campus Regulations” and establishes them on the cloud platform for tax classification.
Term Frequency-Inverse Document Frequency technique can only present one type of measurement and quantitative method and is not capable of presenting diverse selection. Therefore, through the combination of compatibility, Cosine Similarity, Hierarchical Clustering and other techniques, a regulation can produce different results in different compatibility. A wide range of selection can be produced through classification, helping users to find the proper regulations which is related. | en_US |