This paper presents a new density-based clustering algorithm, ST-DBSCAN, which is based on DBSCAN. We propose three marginal extensions to DBSCAN related with the identification of (i) core objects, (ii) noise objects, and (iii) adjacent clusters. In contrast to the existing density-based clustering algorithms, our algorithm has the ability of discovering clusters according to non-spatial, spatial and temporal values of the objects. In this paper, we also present a spatial-temporal data warehouse system designed for storing and clustering a wide range of spatial-temporal data. We show an implementation of our algorithm by using this data warehouse and present the data mining results.

ST-DBSCAN: An algorithm for clustering spatial-temporal data

Text categorization is a significant tool to manage and organize the surging text data. Many text categorization algorithms have been explored in previous literatures, such as KNN, Naive Bayes and Support Vector Machine. KNN text categorization is an effective but less efficient classification method. In this paper, we propose an improved KNN algorithm for text categorization, which builds the classification model by combining constrained one pass clustering algorithm and KNN text categorization. Empirical results on three benchmark corpora show that our algorithm can reduce the text similarity computation substantially and outperform the-state-of-the-art KNN, Naive Bayes and Support Vector Machine classifiers. In addition, the classification model constructed by the proposed algorithm can be updated incrementally, and it has great scalability in many real-word applications.

An improved K-nearest-neighbor algorithm for text categorization

The traditional KNN text classification algorithm used all training samples for classification, so it had a huge number of training samples and a high degree of calculation complexity, and it also didn’t reflect the different importance of different samples. In allusion to the problems mentioned above, an improved KNN text classification algorithm based on clustering center is proposed in this paper. Firstly, the given training sets are compressed and the samples near by the border are deleted, so the multipeak effect of the training sample sets is eliminated. Secondly, the training sample sets of each category are clustered by k-means clustering algorithm, and all cluster centers are taken as the new training samples. Thirdly, a weight value is introduced, which indicates the importance of each training sample according to the number of samples in the cluster that contains this cluster center. Finally, the modified samples are used to accomplish KNN text classification. The simulation results show that the algorithm proposed in this paper can not only effectively reduce the actual number of training samples and lower the calculation complexity, but also improve the accuracy of KNN text classification algorithm.

/pdf/an-improved-knn-text-classification-algorithm-based-on-kndhqe0eff.pdf

An Improved KNN Text Classification Algorithm Based on Clustering

DBSCAN is a well-known density based clustering algorithm capable of discovering arbitrary shaped clusters and eliminating noise data. However, parallelization of Dbscan is challenging as it exhibits an inherent sequential data access order. Moreover, existing parallel implementations adopt a master-slave strategy which can easily cause an unbalanced workload and hence result in low parallel efficiency. We present a new parallel Dbscan algorithm (Pdsdbscan) using graph algorithmic concepts. More specifically, we employ the disjoint-set data structure to break the access sequentiality of Dbscan. In addition, we use a tree-based bottom-up approach to construct the clusters. This yields a better-balanced workload distribution. We implement the algorithm both for shared and for distributed memory. Using data sets containing up to several hundred million high-dimensional points, we show that Pdsdbscan significantly outperforms the master-slave approach, achieving speedups up to 25.97 using 40 cores on shared memory architecture, and speedups up to 5,765 using 8,192 cores on distributed memory architecture.

http://cucis.ece.northwestern.edu/publications/pdf/PatPal12.pdf

A new scalable parallel DBSCAN algorithm using the disjoint-set data structure

New text categorization models using back-propagation neural network (BPNN) and modified back-propagation neural network (MBPNN) are proposed. An efficient feature selection method is used to reduce the dimensionality as well as improve the performance. The basic BPNN learning algorithm has the drawback of slow training speed, so we modify the basic BPNN learning algorithm to accelerate the training speed. The categorization accuracy also has been improved consequently. Traditional word-matching based text categorization system uses vector space model (VSM) to represent the document. However, it needs a high dimensional space to represent the document, and does not take into account the semantic relationship between terms, which can also lead to poor classification accuracy. Latent semantic analysis (LSA) can overcome the problems caused by using statistically derived conceptual indices instead of individual words. It constructs a conceptual vector space in which each term or document is represented as a vector in the space. It not only greatly reduces the dimensionality but also discovers the important associative relationship between terms. We test our categorization models on 20-newsgroup data set, experimental results show that the models using MBPNN outperform than the basic BPNN. And the application of LSA for our system can lead to dramatic dimensionality reduction while achieving good classification results.

Latent semantic analysis for text categorization using neural network

The huge amount of information stored in databases owned by corporations (e.g., retail, financial, telecom) has spurred a tremendous interest in the area of knowledge discovery and data mining. Clustering, in data mining, is a useful technique for discovering interesting data distributions and patterns in the underlying data, and has many application fields, such as statistical data analysis, pattern recognition, image processing, and other business applications. Although researchers have been working on clustering algorithms for decades, and a lot of algorithms for clustering have been developed, there is still no efficient algorithm for clustering very large databases and high dimensional data. As an outstanding representative of clustering algorithms, DBSCAN algorithm shows good performance in spatial data clustering. However, for large spatial databases, DBSCAN requires large volume of memory support and could incur substantial I/O costs because it operates directly on the entire database. In this paper, several approaches are proposed to scale DBSCAN algorithm to large spatial databases. To begin with, a fast DBSCAN algorithm is developed, which considerably speeds up the original DBSCAN algorithm. Then a sampling based DBSCAN algorithm, a partitioning-based DBSCAN algorithm, and a parallel DBSCAN algorithm are introduced consecutively. Following that, based on the above-proposed algorithms, a synthetic algorithm is also given. Finally, some experimental results are given to demonstrate the effectiveness and efficiency of these algorithms.

Approaches for scaling DBSCAN algorithm to large spatial databases

Using Maximum Entropy Model for Chinese Text Categorization

A Simple and Efficient Algorithm to Classify a Large Scale of Texts

Association categorization approach based on frequent patterns has been recently presented, which builds the classification rules according to frequent patterns in various categories and classifies the new text employing these rules. But there are two shortages when the method is applied to classify text data: one is that the method ignores the information about word's frequency in a text; another is that the rule pruning to improve the classification efficiency will lead to obvious descending of accuracy when mass rules are generated. Therefore, a text categorization algorithm based on frequent patterns with term frequency is presented. This study illuminates that the word frequency is helpful for improving the accuracy of the association categorization and the classification rule tree can improve the efficiency of the association classification. The result of experiments shows the performance of association classification is better than three typical text classification methods Bayes, kNN (k nearest neighbor) and SVM (support vector machines), so it is a promising text classification method.

/pdf/text-categorization-based-on-classification-rules-tree-by-3tda9o3gcb.pdf

Hu Yunfa

Papers

Approaches for scaling DBSCAN algorithm to large spatial databases

Using Maximum Entropy Model for Chinese Text Categorization

A Simple and Efficient Algorithm to Classify a Large Scale of Texts

Text Categorization Based on Classification Rules Tree by Frequent Patterns

A Strategy to Class Imbalance Problem for kNN Text Classifier