Text Clustering Algorithm Research Articles

Text Document Featured Clustering and Classification using Fuzzy Logic

Recent years huge amount of information are available in the World Wide web. ‘Text Document Featured Clustering and Classification Using Fuzzy Logic’ aims at obtaining effective clustering of text documents which can simply browsing of large volume of data set using semantic relationship between the words. Text clustering is mainly used for clustering a set of documents based on the user typed key term. ‘A Fuzzy similarity-based algorithm’ automatically classify a group of documents into set of groups with frequent concept. The set of documents and key terms which are entered by the user are pre-processed. The proposed system uses Fuzzy Featured Clustering(FFC) algorithm to identify the semantic relationship of words to create concepts. Text is classified based on features into number of clusters. The frequency of words are weighted by weighted matrix. The relationship between words like synonyms, hypernyms, hyponyms can also be identified. The proposed system is found more accurate, scalable and effective when compared to existing text clustering algorithms. Keywords: Data Mining, Fuzzy Featured Clustering, Text Clustering, Classification, Fuzzy Logic

A Network Decomposition-Based Text Clustering Algorithm for Topic Detection

Text clustering is one of the most popular topic detection techniques. However, the existing text clustering approaches require that each document has to be partitioned to one and only one cluster. This is not reasonable in some cases for there exist some documents which should not used to constitute topics. This paper firstly models a text document set as a network and designs a method for decomposing such a network, and then proposes a truly original text clustering algorithm for topic detection, called a network decomposition-based text clustering algorithm for topic detection (NDTCATD). The proposed algorithm ensures that meaningless documents can not be used to constitute topics. Experimental results show that NDTCATD is much better than bisecting k-means algorithm in terms of overall similarity and average cluster similarity. Therefore the proposed algorithm is reasonable and effective and is especially suitable for topic detection.

Decentralized Probabilistic Text Clustering

Text clustering is an established technique for improving quality in information retrieval, for both centralized and distributed environments. However, traditional text clustering algorithms fail to scale on highly distributed environments, such as peer-to-peer networks. Our algorithm for peer-to-peer clustering achieves high scalability by using a probabilistic approach for assigning documents to clusters. It enables a peer to compare each of its documents only with very few selected clusters, without significant loss of clustering quality. The algorithm offers probabilistic guarantees for the correctness of each document assignment to a cluster. Extensive experimental evaluation with up to 1 million peers and 1 million documents demonstrates the scalability and effectiveness of the algorithm.

High performance genetic algorithm based text clustering using parts of speech and outlier elimination

Among the typical clustering methods, the K-means algorithm plays the most important role in clustering because of its simplicity and efficiency. However, it is sensitive to the initial points and easy to fall into local optimum. In order to avoid this kind of flaw, a patented text clustering algorithm Clustering by Genetic Algorithm Model (CGAM) is revealed in this paper. CGAM constructs the fitness function of genetic algorithm (GA) and convergence criterion for K-means algorithm because GA simulates the natural evolutionary process and deals with a larger search space. To tackle the rich semantics of Chinese texts, CGAM creates an innovative selection method of initial centers of GA and accommodates the contribution of characteristics of different parts of speech. Moreover, the impact of outliers is addressed and treated. Its performance is demonstrated by a series of experiments based on both Reuters-21578 and Chinese text corpus. Experimental results show that the CGAM achieves clustering results better than other GA based K-means algorithms and has been successfully applied to national program of business intelligence system in the context of huge set of contents in both Chinese and English.

Extracting information networks from the blogosphere

We study the problem of automatically extracting information networks formed by recognizable entities as well as relations among them from social media sites. Our approach consists of using state-of-the-art natural language processing tools to identify entities and extract sentences that relate such entities, followed by using text-clustering algorithms to identify the relations within the information network. We propose a new term-weighting scheme that significantly improves on the state-of-the-art in the task of relation extraction, both when used in conjunction with the standard tf ċ idf scheme and also when used as a pruning filter. We describe an effective method for identifying benchmarks for open information extraction that relies on a curated online database that is comparable to the hand-crafted evaluation datasets in the literature. From this benchmark, we derive a much larger dataset which mimics realistic conditions for the task of open information extraction. We report on extensive experiments on both datasets, which not only shed light on the accuracy levels achieved by state-of-the-art open information extraction tools, but also on how to tune such tools for better results.

Short Text Clustering Algorithm with Feature Keyword Expansion

To solve the problem of sparse keywords and similarity drift in short text segments, this paper proposes short text clustering algorithm with feature keyword expansion (STCAFKE). The method can realize short text clustering by expanding feature keyword based on HowNet and combining K-means algorithm and density algorithm. It may add the number of text keyword with feature keyword expansion and increase text semantic features to realize short text clustering. Experimental results show that this algorithm has increased the short text clustering quality on precision and recall.

Document Clustering Based on Semi-Supervised Term Clustering

The study is conducted to propose a multi-step feature (term) selection process and in semi-supervised fashion, provide initial centers for term clusters. Then utilize the fuzzy c-means (FCM) clustering algorithm for clustering terms. Finally assign each of documents to closest associated term clusters. While most text clustering algorithms directly use documents for clustering, we propose to first group the terms using FCM algorithm and then cluster documents based on terms clusters. We evaluate effectiveness of our technique on several standard text collections and compare our results with the some classical text clustering algorithms.

An Online Hot Topics Detection Approach Using the Improved Ant Colony Text Clustering Algorithm

Recently, with an increasing number of major events spreading all over the Internet, the research for online hot topics detection system has been paid more and more attention. In this paper, we proposed an unsupervised and efficient hot topics detection approach, which is based on an improved ant colony text clustering (IACTC) algorithm. In view of the deficiencies of the basic ant colony text clustering (BACTC) method, the inverse tangent function is introduced into the proposed algorithm as a novel probability function, which could be flexibly adjusted. Meanwhile, in order to reduce ants’ blinding movement, we add the memory organ to each ant, and we develop the concept of adaptive moving range. Additionally, we design a practical system framework with technical details to solve the problems involving in the real-time hot topics detection system (e.g., information collecting, web preprocessing and text processing). This system is also used to evaluate our proposed method .The experimental results show that our approach has numerous advantages and achieves satisfying results.

A Novel Method For Document Clustering Using Ant-fuzzy Algorithm

Availability of large full-text document collection in electronic forms has created a need for tools techniques that assist users in organization. Document clustering is one of the popular methods used for this purpose. Ant-based text clustering is a promising technique that has attracted great research attention. This paper attempts to improve the standard ant-based text-clustering algorithm. The ant behavior model is modified to pursue better algorithmic performance. In this paper, a hybrid approach based on Ant clustering and Fuzzy clustering methods is used. First ant based clustering is used for creating raw and imprecise clusters and then these clusters are refined by means of fuzzy C-Mean (FCM) algorithm. For large datasets these two stages does not suffice and many homogenous small clusters are formed. Thus more iteration of these two stages is usually required and clusters from previous iterations are used as a building block in the following iterations to build finer and larger clusters. The proposed algorithm is tested with a sample set of documents excerpted from the Reuters-21578 corpus and the experiment results partly indicate that the proposed algorithm perform better than the standard ant-based text-clustering algorithm and the k-means algorithm.

A New Text Clustering Algorithm Based on Improved K_means

Text clustering is one of the difficult and hot research fields in the internet search engine research. A new text clustering algorithm is presented based on K-means and Self-Organizing Model (SOM). Firstly, texts are preprocessed to satisfy succeed process requirement. Secondly, the paper improves selection of initial cluster centers and cluster seed selection methods of K-means to improve the deficiency of K-means algorithm that the K-means algorithm is very sensitive to the initial cluster center and the isolated point text. Thirdly the advantages of k-means and SOM are combined to a new model to cluster text in the paper. Finally the experimental results indicate that the improved algorithm has a higher accuracy compared with the original algorithm, and has a better stability.

A Novel Text Clustering Algorithm Based on Feature Weighting Distance and Soft Subspace Learning

A Novel Text Clustering Algorithm Based on Feature Weighting Distance and Soft Subspace Learning

ML-CLUBAS: A Multi Label Bug Classification Algorithm

In this paper, a multi label variant of CLUBAS [1] algorithm, ML-CLUBAS (Multi Label-Classification of software Bugs Using Bug Attribute Similarity) is presented. CLUBAS is a hybrid algorithm, and is designed by using text clustering, frequent term calculations and taxonomic terms mapping techniques, and is an example of classification using clustering technique. CLUBAS is a single label algorithm, where one bug cluster is exactly mapped to a single bug category. However a bug cluster can be mapped into the more than one bug category in case of cluster label matches with the more than one category term, for this purpose ML-CLUBAS a multi label variant of CLUBAS is presented in this work. The designed algorithm is evaluated using the performance parameters F-measures and accuracy, number of clusters and purity. These parameters are compared with the CLUBAS and other multi label text clustering algorithms.

Functional grouping of natural language requirements for assistance in architectural software design

Modern software systems are becoming larger and more complex every day. One of the most challenging steps for designing a good architecture for a certain piece of software is the analysis of requirements, usually written in natural language by engineers not familiar with specific design formalisms. The main problem related to this task is the conceptual gap existing between low-level requirements and higher views of the system decomposing its functionality. In this paper, we introduce an approach for mining and grouping functionality from textual descriptions of requirements using text mining techniques aiming at helping software designers with this complex and time-consuming task. The knowledge discovered starting from informally written requirements using a combination of natural language processing (NLP) and text clustering algorithms can be then easily mapped into design concerns of a possible architecture for the system. Experimental validation in three case studies suggests a great potential of the proposed approach for providing assistance to software designers during early stages of the software development process.

Improved text clustering algorithm of probabilistic latent with semantic analysis

Trained by the Expectation Maximization(EM) algorithm,whose model parameters are randomly initialized,the performance of Probabilistic Latent Semantic Analysis(PLSA) model is quite dependent on the initialization of the model,and the result of iteration is not a global maximum,but a local one.The authors derived probabilities from Latent Semantic Analysis(LSA),and then used it to initialize the parameters of PLSA model in documents clustering.The improved PLSA could effectively solve the puzzle of random initializing of EM.It is shown that the improved algorithm has a distinct improvement in Normalized Mutual Information(NMI) and accuracy.

Text Clustering with Seeds Affinity Propagation

Based on an effective clustering algorithm-Affinity Propagation (AP)-we present in this paper a novel semisupervised text clustering algorithm, called Seeds Affinity Propagation (SAP). There are two main contributions in our approach: 1) a new similarity metric that captures the structural information of texts, and 2) a novel seed construction method to improve the semisupervised clustering process. To study the performance of the new algorithm, we applied it to the benchmark data set Reuters-21578 and compared it to two state-of-the-art clustering algorithms, namely, k-means algorithm and the original AP algorithm. Furthermore, we have analyzed the individual impact of the two proposed contributions. Results show that the proposed similarity metric is more effective in text clustering (F-measures ca. 21 percent higher than in the AP algorithm) and the proposed semisupervised strategy achieves both better clustering results and faster convergence (using only 76 percent iterations of the original AP). The complete SAP algorithm obtains higher F-measure (ca. 40 percent improvement over k-means and AP) and lower entropy (ca. 28 percent decrease over k-means and AP), improves significantly clustering execution time (20 times faster) in respect that k-means, and provides enhanced robustness compared with all other methods.

A Text Hybrid Clustering Algorithm Based on HowNet Semantics

Many existing text clustering algorithms overlook the semantic information between words and so they possess a lower accuracy of text similarity computation. A new text hybrid clustering algorithm (HCA) based on HowNet semantics has been proposed in this paper. It calculates the semantic similarity of words by using the words’ semantic concept description in HowNet and then combines it with the method of maximum weight matching of bipartite graph to calculate a semantic-based text similarity. Based on the new text similarity and by combining an improved genetic algorithm with k-medoids algorithm, HCA has been designed. The comparative experiments show that: 1) compared with two existing traditional clustering algorithms, HCA can get better quality and 2) when their text cosine similarity is replaced with the new semantic-based text similarity, all the qualities of the three clustering algorithms can be improved significantly.

Text Clustering Based on the User Search Intention

This paper presents a novel algorithm of Text Clustering. With the popularity of the Internet, text information on the web shows explosive growth trend. Text Clustering technology as a method of unsupervised machine learning, which does not need the training process and pre-manual tagging, so Text Clustering is an effective way for dealing with massive text messages. The traditional Text Clustering is based on the content of the article, and they think that the articles which belong to the same class have the greater similarity. In this paper, we extracted label word from the summary information returned by search engine. Then did hierarchical clustering based on the text feature of the label word. Experiment shows that the algorithm is feasible.

A Cluster Tree Method For Text Categorization

The decision tree is a flexible and useful classification tool. But on the data with high dimensionality, it meets problems. For most of current decision tree algorithms, when splitting a node of a tree, only the “best” one feature is selected and used. Since more features are ignored, the classification accuracy is not high. To solve the problem, this paper uses a cluster tree for text categorization. Unlike familiar decision trees (e.g. CART, C4.5), clustering results are used as the splitting rule and more features are considered. Obviously, the used clustering algorithm is an very important to the cluster tree. For better performance, a text clustering algorithm is proposed to enhance the cluster tree. Experiments show that the cluster tree solves the high-dimensionality problem and outperforms C4.5 and CART on text data. Sometimes, it may do better than LibSVM, which may be the most powerful tool for text categorization.

K-means text clustering algorithm based on density and nearest neighbor

The selection of initial focal point has great influence on the clustering results of traditional K-means algorithm,for it tends to get a local optimal solution when inappropriately assigned.In view of this issue,initial algorithm that can generate the initial cluster center was proposed,through introducing the density and nearest neighbor idea.These selected centers were used for K-means algorithm;a better text clustering algorithm called DN-K-means was put forward.The results of experiments indicate that the algorithm can lead to results with high and steady clustering quality.

A New Agglomerative Hierarchical Clustering Algorithm Implementation based on the Map Reduce Framework

Text clustering is one of the difficult and hot research fields in the text mining research. Combing Map Reduce framework and the neuron initialization method of VPSOM (vector pressing SelfOrganizing Model) algorithm, a new text clustering algorithm is presented. It divides the large text vector dataset into data blocks, each of which then processed in different distributed data node of Map Reduce framework with agglomerative hierarchical clustering algorithm. The experiment results indicate that the improved algorithm has a higher efficiency and a better accuracy.