In order to improve the clustering results and select in the results, the ontology semantic is combined with document clustering. A new document clustering algorithm based WordNet in the phrase of document processing ...In order to improve the clustering results and select in the results, the ontology semantic is combined with document clustering. A new document clustering algorithm based WordNet in the phrase of document processing is proposed. First, every word vector by new entities is extended after the documents are represented by tf-idf. Then the feature extracting algorithm is applied for the documents. Finally, the algorithm of ontology aggregation clustering （OAC） is proposed to improve the result of document clustering. Experiments are based on the data set of Reuters 20 News Group, and experimental results are compared with the results obtained by mutual information（MI）. The conclusion draws that the proposed algorithm of document clustering based on ontology is better than the other existed clustering algorithms such as MNB, CLUTO, co-clustering, etc.展开更多
To improve the accuracy of text clustering, fuzzy c-means clustering based on topic concept sub-space （TCS2FCM） is introduced for classifying texts. Five evaluation functions are combined to extract key phrases. Con...To improve the accuracy of text clustering, fuzzy c-means clustering based on topic concept sub-space （TCS2FCM） is introduced for classifying texts. Five evaluation functions are combined to extract key phrases. Concept phrases, as well as the descriptions of final clusters, are presented using WordNet origin from key phrases. Initial centers and membership matrix are the most important factors affecting clustering performance. Orthogonal concept topic sub-spaces are built with the topic concept phrases representing topics of the texts and the initialization of centers and the membership matrix depend on the concept vectors in sub-spaces. The results show that, different from random initialization of traditional fuzzy c-means clustering, the initialization related to text content contributions can improve clustering precision.展开更多
A method that combines category-based and keyword-based concepts for a better information retrieval system is introduced. To improve document clustering, a document similarity measure based on cosine vector and keywor...A method that combines category-based and keyword-based concepts for a better information retrieval system is introduced. To improve document clustering, a document similarity measure based on cosine vector and keywords frequency in documents is proposed, but also with an input ontology. The ontology is domain specific and includes a list of keywords organized by degree of importance to the categories of the ontology, and by means of semantic knowledge, the ontology can improve the effects of document similarity measure and feedback of information retrieval systems. Two approaches to evaluating the performance of this similarity measure and the comparison with standard cosine vector similarity measure are also described.展开更多
基金The National Natural Science Foundation of China(No.60373099),the Natural Science Foundation for Young Scholars of Northeast Normal University (No.20061005)
文摘In order to improve the clustering results and select in the results, the ontology semantic is combined with document clustering. A new document clustering algorithm based WordNet in the phrase of document processing is proposed. First, every word vector by new entities is extended after the documents are represented by tf-idf. Then the feature extracting algorithm is applied for the documents. Finally, the algorithm of ontology aggregation clustering （OAC） is proposed to improve the result of document clustering. Experiments are based on the data set of Reuters 20 News Group, and experimental results are compared with the results obtained by mutual information（MI）. The conclusion draws that the proposed algorithm of document clustering based on ontology is better than the other existed clustering algorithms such as MNB, CLUTO, co-clustering, etc.
基金The National Natural Science Foundation of China(No60672056)Open Fund of MOE-MS Key Laboratory of Multime-dia Computing and Communication(No06120809)
文摘To improve the accuracy of text clustering, fuzzy c-means clustering based on topic concept sub-space （TCS2FCM） is introduced for classifying texts. Five evaluation functions are combined to extract key phrases. Concept phrases, as well as the descriptions of final clusters, are presented using WordNet origin from key phrases. Initial centers and membership matrix are the most important factors affecting clustering performance. Orthogonal concept topic sub-spaces are built with the topic concept phrases representing topics of the texts and the initialization of centers and the membership matrix depend on the concept vectors in sub-spaces. The results show that, different from random initialization of traditional fuzzy c-means clustering, the initialization related to text content contributions can improve clustering precision.
基金The Young Teachers Scientific Research Foundation (YTSRF) of Nanjing University of Science and Technology in the Year of2005-2006.
文摘A method that combines category-based and keyword-based concepts for a better information retrieval system is introduced. To improve document clustering, a document similarity measure based on cosine vector and keywords frequency in documents is proposed, but also with an input ontology. The ontology is domain specific and includes a list of keywords organized by degree of importance to the categories of the ontology, and by means of semantic knowledge, the ontology can improve the effects of document similarity measure and feedback of information retrieval systems. Two approaches to evaluating the performance of this similarity measure and the comparison with standard cosine vector similarity measure are also described.
